♻ ☆ Tra-MoE: Learning Trajectory Prediction Model from Multiple Domains for Adaptive Policy Conditioning CVPR 2025

Jiange Yang, Haoyi Zhu, Yating Wang, Gangshan Wu, Tong He, Limin Wang

Learning from multiple domains is a primary factor that influences the generalization of a single unified robot system. In this paper, we aim to learn the trajectory prediction model by using broad out-of-domain data to improve its performance and generalization ability. Trajectory model is designed to predict any-point trajectories in the current frame given an instruction and can provide detailed control guidance for robotic policy learning. To handle the diverse out-of-domain data distribution, we propose a sparsely-gated MoE (\textbf{Top-1} gating strategy) architecture for trajectory model, coined as \textbf{Tra-MoE}. The sparse activation design enables good balance between parameter cooperation and specialization, effectively benefiting from large-scale out-of-domain data while maintaining constant FLOPs per token. In addition, we further introduce an adaptive policy conditioning technique by learning 2D mask representations for predicted trajectories, which is explicitly aligned with image observations to guide action prediction more flexibly. We perform extensive experiments on both simulation and real-world scenarios to verify the effectiveness of Tra-MoE and adaptive policy conditioning technique. We also conduct a comprehensive empirical study to train Tra-MoE, demonstrating that our Tra-MoE consistently exhibits superior performance compared to the dense baseline model, even when the latter is scaled to match Tra-MoE's parameter count.

comment: Accepted to CVPR 2025. Code Page: https://github.com/MCG-NJU/Tra-MoE

♻ ☆ ActiveGAMER: Active GAussian Mapping through Efficient Rendering CVPR2025

Liyan Chen, Huangying Zhan, Kevin Chen, Xiangyu Xu, Qingan Yan, Changjiang Cai, Yi Xu

We introduce ActiveGAMER, an active mapping system that utilizes 3D Gaussian Splatting (3DGS) to achieve high-quality, real-time scene mapping and exploration. Unlike traditional NeRF-based methods, which are computationally demanding and restrict active mapping performance, our approach leverages the efficient rendering capabilities of 3DGS, allowing effective and efficient exploration in complex environments. The core of our system is a rendering-based information gain module that dynamically identifies the most informative viewpoints for next-best-view planning, enhancing both geometric and photometric reconstruction accuracy. ActiveGAMER also integrates a carefully balanced framework, combining coarse-to-fine exploration, post-refinement, and a global-local keyframe selection strategy to maximize reconstruction completeness and fidelity. Our system autonomously explores and reconstructs environments with state-of-the-art geometric and photometric accuracy and completeness, significantly surpassing existing approaches in both aspects. Extensive evaluations on benchmark datasets such as Replica and MP3D highlight ActiveGAMER's effectiveness in active mapping tasks.

comment: Accepted to CVPR2025

♻ ☆ TelePreview: A User-Friendly Teleoperation System with Virtual Arm Assistance for Enhanced Effectiveness

Jingxiang Guo, Jiayu Luo, Zhenyu Wei, Yiwen Hou, Zhixuan Xu, Xiaoyi Lin, Chongkai Gao, Lin Shao

Teleoperation provides an effective way to collect robot data, which is crucial for learning from demonstrations. In this field, teleoperation faces several key challenges: user-friendliness for new users, safety assurance, and transferability across different platforms. While collecting real robot dexterous manipulation data by teleoperation to train robots has shown impressive results on diverse tasks, due to the morphological differences between human and robot hands, it is not only hard for new users to understand the action mapping but also raises potential safety concerns during operation. To address these limitations, we introduce TelePreview. This teleoperation system offers real-time visual feedback on robot actions based on human user inputs, with a total hardware cost of less than $1,000. TelePreview allows the user to see a virtual robot that represents the outcome of the user's next movement. By enabling flexible switching between command visualization and actual execution, this system helps new users learn how to demonstrate quickly and safely. We demonstrate that it outperforms other teleoperation systems across five tasks, emphasize its ease of use, and highlight its straightforward deployment across diverse robotic platforms. We release our code and a deployment document on our website https://nus-lins-lab.github.io/telepreview-web/.

comment: In submission

♻ ☆ One Policy to Run Them All: an End-to-end Learning Approach to Multi-Embodiment Locomotion

Nico Bohlinger, Grzegorz Czechmanowski, Maciej Krupka, Piotr Kicki, Krzysztof Walas, Jan Peters, Davide Tateo

Deep Reinforcement Learning techniques are achieving state-of-the-art results in robust legged locomotion. While there exists a wide variety of legged platforms such as quadruped, humanoids, and hexapods, the field is still missing a single learning framework that can control all these different embodiments easily and effectively and possibly transfer, zero or few-shot, to unseen robot embodiments. We introduce URMA, the Unified Robot Morphology Architecture, to close this gap. Our framework brings the end-to-end Multi-Task Reinforcement Learning approach to the realm of legged robots, enabling the learned policy to control any type of robot morphology. The key idea of our method is to allow the network to learn an abstract locomotion controller that can be seamlessly shared between embodiments thanks to our morphology-agnostic encoders and decoders. This flexible architecture can be seen as a potential first step in building a foundation model for legged robot locomotion. Our experiments show that URMA can learn a locomotion policy on multiple embodiments that can be easily transferred to unseen robot platforms in simulation and the real world.

♻ ☆ A Digital Twin for Telesurgery under Intermittent Communication

Junxiang Wang, Juan Antonio Barragan, Hisashi Ishida, Jingkai Guo, Yu-Chun Ku, Peter Kazanzides

Telesurgery is an effective way to deliver service from expert surgeons to areas without immediate access to specialized resources. However, many of these areas, such as rural districts or battlefields, might be subject to different problems in communication, especially latency and intermittent periods of communication outage. This challenge motivates the use of a digital twin for the surgical system, where a simulation would mirror the robot hardware and surgical environment in the real world. The surgeon would then be able to interact with the digital twin during communication outage, followed by a recovery strategy on the real robot upon reestablishing communication. This paper builds the digital twin for the da Vinci surgical robot, with a buffering and replay strategy that reduces the mean task completion time by 23% when compared to the baseline, for a peg transfer task subject to intermittent communication outage. The relevant code can be found here: https://github.com/LCSR-CIIS/dvrk_digital_twin_teleoperation.

comment: 7 pages, 5 figures. To be published in 2025 International Symposium on Medical Robotics (ISMR)

♻ ☆ Hierarchical Procedural Framework for Low-latency Robot-Assisted Hand-Object Interaction

Mingqi Yuan, Huijiang Wang, Kai-Fung Chu, Fumiya Iida, Bo Li, Wenjun Zeng

Advances in robotics have been driving the development of human-robot interaction (HRI) technologies. However, accurately perceiving human actions and achieving adaptive control remains a challenge in facilitating seamless coordination between human and robotic movements. In this paper, we propose a hierarchical procedural framework to enable dynamic robot-assisted hand-object interaction. An open-loop hierarchy leverages the computer vision (CV)-based 3D reconstruction of the human hand, based on which motion primitives have been designed to translate hand motions into robotic actions. The low-level coordination hierarchy fine-tunes the robot's action by using the continuously updated 3D hand models. Experimental validation demonstrates the effectiveness of the hierarchical control architecture. The adaptive coordination between human and robot behavior has achieved a delay of $\leq 0.3$ seconds in the tele-interaction scenario. A case study of ring-wearing tasks indicates the potential application of this work in assistive technologies such as healthcare and manufacturing.

comment: 6 pages, 5 figures

♻ ☆ DELTA: Decomposed Efficient Long-Term Robot Task Planning using Large Language Models ICRA 2025

Yuchen Liu, Luigi Palmieri, Sebastian Koch, Ilche Georgievski, Marco Aiello

Recent advancements in Large Language Models (LLMs) have sparked a revolution across many research fields. In robotics, the integration of common-sense knowledge from LLMs into task and motion planning has drastically advanced the field by unlocking unprecedented levels of context awareness. Despite their vast collection of knowledge, large language models may generate infeasible plans due to hallucinations or missing domain information. To address these challenges and improve plan feasibility and computational efficiency, we introduce DELTA, a novel LLM-informed task planning approach. By using scene graphs as environment representations within LLMs, DELTA achieves rapid generation of precise planning problem descriptions. To enhance planning performance, DELTA decomposes long-term task goals with LLMs into an autoregressive sequence of sub-goals, enabling automated task planners to efficiently solve complex problems. In our extensive evaluation, we show that DELTA enables an efficient and fully automatic task planning pipeline, achieving higher planning success rates and significantly shorter planning times compared to the state of the art. Project webpage: https://delta-llm.github.io/

comment: Accepted at ICRA 2025

♻ ☆ RedMotion: Motion Prediction via Redundancy Reduction

Royden Wagner, Omer Sahin Tas, Marvin Klemp, Carlos Fernandez, Christoph Stiller

We introduce RedMotion, a transformer model for motion prediction in self-driving vehicles that learns environment representations via redundancy reduction. Our first type of redundancy reduction is induced by an internal transformer decoder and reduces a variable-sized set of local road environment tokens, representing road graphs and agent data, to a fixed-sized global embedding. The second type of redundancy reduction is obtained by self-supervised learning and applies the redundancy reduction principle to embeddings generated from augmented views of road environments. Our experiments reveal that our representation learning approach outperforms PreTraM, Traj-MAE, and GraphDINO in a semi-supervised setting. Moreover, RedMotion achieves competitive results compared to HPTR or MTR++ in the Waymo Motion Prediction Challenge. Our open-source implementation is available at: https://github.com/kit-mrt/future-motion

comment: TMLR published version

♻ ☆ A formal implementation of Behavior Trees to act in robotics

Felix Ingrand

Behavior Trees (BT) are becoming quite popular as an Acting component of autonomous robotic systems. We propose to define a formal semantics to BT by translating them to a formal language which enables us to perform verification of programs written with BT, as well as runtime verification while these BT execute. This allows us to formally verify BT correctness without requiring BT programmers to master formal languages and without compromising BT most valuable features: modularity, flexibility and reusability. We present the formal framework we use: Fiacre, its language and the produced TTS model; Tina, its model checking tools and Hippo, its runtime verification engine. We then show how the translation from BT to Fiacre is automatically done, the type of formal LTL and CTL properties we can check offline and how to execute the formal model online in place of a regular BT engine. We illustrate our approach on two robotics applications, and show how BT can be extended with state variables, eval nodes, node evaluation results and benefit of other features available in the Fiacre formal framework (e.g., time).

♻ ☆ AVOCADO: Adaptive Optimal Collision Avoidance driven by Opinion

Diego Martinez-Baselga, Eduardo Sebastián, Eduardo Montijano, Luis Riazuelo, Carlos Sagüés, Luis Montano

We present AVOCADO (AdaptiVe Optimal Collision Avoidance Driven by Opinion), a novel navigation approach to address holonomic robot collision avoidance when the robot does not know how cooperative the other agents in the environment are. AVOCADO departs from a Velocity Obstacle's (VO) formulation akin to the Optimal Reciprocal Collision Avoidance method. However, instead of assuming reciprocity, it poses an adaptive control problem to adapt to the cooperation level of other robots and agents in real time. This is achieved through a novel nonlinear opinion dynamics design that relies solely on sensor observations. As a by-product, we leverage tools from the opinion dynamics formulation to naturally avoid the deadlocks in geometrically symmetric scenarios that typically suffer VO-based planners. Extensive numerical simulations show that AVOCADO surpasses existing motion planners in mixed cooperative/non-cooperative navigation environments in terms of success rate, time to goal and computational time. In addition, we conduct multiple real experiments that verify that AVOCADO is able to avoid collisions in environments crowded with other robots and humans.

comment: This paper is published at IEEE Transactions on Robotics under DOI 10.1109/TRO.2025.3552350

♻ ☆ A Graph-to-Text Approach to Knowledge-Grounded Response Generation in Human-Robot Interaction

Nicholas Thomas Walker, Stefan Ultes, Pierre Lison

Knowledge graphs are often used to represent structured information in a flexible and efficient manner, but their use in situated dialogue remains under-explored. This paper presents a novel conversational model for human--robot interaction that rests upon a graph-based representation of the dialogue state. The knowledge graph representing the dialogue state is continuously updated with new observations from the robot sensors, including linguistic, situated and multimodal inputs, and is further enriched by other modules, in particular for spatial understanding. The neural conversational model employed to respond to user utterances relies on a simple but effective graph-to-text mechanism that traverses the dialogue state graph and converts the traversals into a natural language form. This conversion of the state graph into text is performed using a set of parameterized functions, and the values for those parameters are optimized based on a small set of Wizard-of-Oz interactions. After this conversion, the text representation of the dialogue state graph is included as part of the prompt of a large language model used to decode the agent response. The proposed approach is empirically evaluated through a user study with a humanoid robot that acts as conversation partner to evaluate the impact of the graph-to-text mechanism on the response generation. After moving a robot along a tour of an indoor environment, participants interacted with the robot using spoken dialogue and evaluated how well the robot was able to answer questions about what the robot observed during the tour. User scores show a statistically significant improvement in the perceived factuality of the robot responses when the graph-to-text approach is employed, compared to a baseline using inputs structured as semantic triples.

comment: Submitted to Dialogue & Discourse 2023

♻ ☆ Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation

Han Xue, Jieji Ren, Wendi Chen, Gu Zhang, Yuan Fang, Guoying Gu, Huazhe Xu, Cewu Lu

Humans can accomplish complex contact-rich tasks using vision and touch, with highly reactive capabilities such as quick adjustments to environmental changes and adaptive control of contact forces; however, this remains challenging for robots. Existing visual imitation learning (IL) approaches rely on action chunking to model complex behaviors, which lacks the ability to respond instantly to real-time tactile feedback during the chunk execution. Furthermore, most teleoperation systems struggle to provide fine-grained tactile / force feedback, which limits the range of tasks that can be performed. To address these challenges, we introduce TactAR, a low-cost teleoperation system that provides real-time tactile feedback through Augmented Reality (AR), along with Reactive Diffusion Policy (RDP), a novel slow-fast visual-tactile imitation learning algorithm for learning contact-rich manipulation skills. RDP employs a two-level hierarchy: (1) a slow latent diffusion policy for predicting high-level action chunks in latent space at low frequency, (2) a fast asymmetric tokenizer for closed-loop tactile feedback control at high frequency. This design enables both complex trajectory modeling and quick reactive behavior within a unified framework. Through extensive evaluation across three challenging contact-rich tasks, RDP significantly improves performance compared to state-of-the-art visual IL baselines through rapid response to tactile / force feedback. Furthermore, experiments show that RDP is applicable across different tactile / force sensors. Code and videos are available on https://reactive-diffusion-policy.github.io.

♻ ☆ AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors ICLR 2025

Ruoxuan Feng, Jiangyu Hu, Wenke Xia, Tianci Gao, Ao Shen, Yuhao Sun, Bin Fang, Di Hu

Visuo-tactile sensors aim to emulate human tactile perception, enabling robots to precisely understand and manipulate objects. Over time, numerous meticulously designed visuo-tactile sensors have been integrated into robotic systems, aiding in completing various tasks. However, the distinct data characteristics of these low-standardized visuo-tactile sensors hinder the establishment of a powerful tactile perception system. We consider that the key to addressing this issue lies in learning unified multi-sensor representations, thereby integrating the sensors and promoting tactile knowledge transfer between them. To achieve unified representation of this nature, we introduce TacQuad, an aligned multi-modal multi-sensor tactile dataset from four different visuo-tactile sensors, which enables the explicit integration of various sensors. Recognizing that humans perceive the physical environment by acquiring diverse tactile information such as texture and pressure changes, we further propose to learn unified multi-sensor representations from both static and dynamic perspectives. By integrating tactile images and videos, we present AnyTouch, a unified static-dynamic multi-sensor representation learning framework with a multi-level structure, aimed at both enhancing comprehensive perceptual abilities and enabling effective cross-sensor transfer. This multi-level architecture captures pixel-level details from tactile data via masked modeling and enhances perception and transferability by learning semantic-level sensor-agnostic features through multi-modal alignment and cross-sensor matching. We provide a comprehensive analysis of multi-sensor transferability, and validate our method on various datasets and in the real-world pouring task. Experimental results show that our method outperforms existing methods, exhibits outstanding static and dynamic perception capabilities across various sensors.

comment: Accepted by ICLR 2025

♻ ☆ VET: A Visual-Electronic Tactile System for Immersive Human-Machine Interaction

Cong Zhang, Yisheng Yang, Shilong Mu, Chuqiao Lyu, Shoujie Li, Xinyue Chai, Wenbo Ding

In the pursuit of deeper immersion in human-machine interaction, achieving higher-dimensional tactile input and output on a single interface has become a key research focus. This study introduces the Visual-Electronic Tactile (VET) System, which builds upon vision-based tactile sensors (VBTS) and integrates electrical stimulation feedback to enable bidirectional tactile communication. We propose and implement a system framework that seamlessly integrates an electrical stimulation film with VBTS using a screen-printing preparation process, eliminating interference from traditional methods. While VBTS captures multi-dimensional input through visuotactile signals, electrical stimulation feedback directly stimulates neural pathways, preventing interference with visuotactile information. The potential of the VET system is demonstrated through experiments on finger electrical stimulation sensitivity zones, as well as applications in interactive gaming and robotic arm teleoperation. This system paves the way for new advancements in bidirectional tactile interaction and its broader applications.

♻ ☆ Temporal and Semantic Evaluation Metrics for Foundation Models in Post-Hoc Analysis of Robotic Sub-tasks IROS 2024

Jonathan Salfity, Selma Wanna, Minkyu Choi, Mitch Pryor

Recent works in Task and Motion Planning (TAMP) show that training control policies on language-supervised robot trajectories with quality labeled data markedly improves agent task success rates. However, the scarcity of such data presents a significant hurdle to extending these methods to general use cases. To address this concern, we present an automated framework to decompose trajectory data into temporally bounded and natural language-based descriptive sub-tasks by leveraging recent prompting strategies for Foundation Models (FMs) including both Large Language Models (LLMs) and Vision Language Models (VLMs). Our framework provides both time-based and language-based descriptions for lower-level sub-tasks that comprise full trajectories. To rigorously evaluate the quality of our automatic labeling framework, we contribute an algorithm SIMILARITY to produce two novel metrics, temporal similarity and semantic similarity. The metrics measure the temporal alignment and semantic fidelity of language descriptions between two sub-task decompositions, namely an FM sub-task decomposition prediction and a ground-truth sub-task decomposition. We present scores for temporal similarity and semantic similarity above 90%, compared to 30% of a randomized baseline, for multiple robotic environments, demonstrating the effectiveness of our proposed framework. Our results enable building diverse, large-scale, language-supervised datasets for improved robotic TAMP.

comment: 8 pages, 3 figures. IROS 2024 Submission

♻ ☆ Scalable Real2Sim: Physics-Aware Asset Generation Via Robotic Pick-and-Place Setups

Nicholas Pfaff, Evelyn Fu, Jeremy Binagia, Phillip Isola, Russ Tedrake

Simulating object dynamics from real-world perception shows great promise for digital twins and robotic manipulation but often demands labor-intensive measurements and expertise. We present a fully automated Real2Sim pipeline that generates simulation-ready assets for real-world objects through robotic interaction. Using only a robot's joint torque sensors and an external camera, the pipeline identifies visual geometry, collision geometry, and physical properties such as inertial parameters. Our approach introduces a general method for extracting high-quality, object-centric meshes from photometric reconstruction techniques (e.g., NeRF, Gaussian Splatting) by employing alpha-transparent training while explicitly distinguishing foreground occlusions from background subtraction. We validate the full pipeline through extensive experiments, demonstrating its effectiveness across diverse objects. By eliminating the need for manual intervention or environment modifications, our pipeline can be integrated directly into existing pick-and-place setups, enabling scalable and efficient dataset creation. Project page (with code and data): https://scalable-real2sim.github.io/.

comment: Website: https://scalable-real2sim.github.io/

♻ ☆ RG-Attn: Radian Glue Attention for Multi-modality Multi-agent Cooperative Perception

Lantao Li, Kang Yang, Wenqi Zhang, Xiaoxue Wang, Chen Sun

Cooperative perception offers an optimal solution to overcome the perception limitations of single-agent systems by leveraging Vehicle-to-Everything (V2X) communication for data sharing and fusion across multiple agents. However, most existing approaches focus on single-modality data exchange, limiting the potential of both homogeneous and heterogeneous fusion across agents. This overlooks the opportunity to utilize multi-modality data per agent, restricting the system's performance. In the automotive industry, manufacturers adopt diverse sensor configurations, resulting in heterogeneous combinations of sensor modalities across agents. To harness the potential of every possible data source for optimal performance, we design a robust LiDAR and camera cross-modality fusion module, Radian-Glue-Attention (RG-Attn), applicable to both intra-agent cross-modality fusion and inter-agent cross-modality fusion scenarios, owing to the convenient coordinate conversion by transformation matrix and the unified sampling/inversion mechanism. We also propose two different architectures, named Paint-To-Puzzle (PTP) and Co-Sketching-Co-Coloring (CoS-CoCo), for conducting cooperative perception. PTP aims for maximum precision performance and achieves smaller data packet size by limiting cross-agent fusion to a single instance, but requiring all participants to be equipped with LiDAR. In contrast, CoS-CoCo supports agents with any configuration-LiDAR-only, camera-only, or LiDAR-camera-both, presenting more generalization ability. Our approach achieves state-of-the-art (SOTA) performance on both real and simulated cooperative perception datasets. The code is now available at GitHub.

♻ ☆ Whole-Body Dynamic Throwing with Legged Manipulators

Humphrey Munn, Brendan Tidd, Peter Böhm, Marcus Gallagher, David Howard

Throwing with a legged robot involves precise coordination of object manipulation and locomotion - crucial for advanced real-world interactions. Most research focuses on either manipulation or locomotion, with minimal exploration of tasks requiring both. This work investigates leveraging all available motors (full-body) over arm-only throwing in legged manipulators. We frame the task as a deep reinforcement learning (RL) objective, optimising throwing accuracy towards any user-commanded target destination and the robot's stability. Evaluations on a humanoid and an armed quadruped in simulation show that full-body throwing improves range, accuracy, and stability by exploiting body momentum, counter-balancing, and full-body dynamics. We introduce an optimised adaptive curriculum to balance throwing accuracy and stability, along with a tailored RL environment setup for efficient learning in sparse-reward conditions. Unlike prior work, our approach generalises to targets in 3D space. We transfer our learned controllers from simulation to a real humanoid platform.

♻ ☆ Analysis and Optimization of Robustness in Multiplex Flow Networks Against Cascading Failures

Orkun İrsoy, Osman Yağan

Networked systems are susceptible to cascading failures, where the failure of an initial set of nodes propagates through the network, often leading to system-wide failures. In this work, we propose a multiplex flow network model to study robustness against cascading failures triggered by random failures. The model is inspired by systems where nodes carry or support multiple types of flows, and failures result in the redistribution of flows within the same layer rather than between layers. To represent different types of interdependencies between the layers of the multiplex network, we define two cases of failure conditions: layer-independent overload and layer-influenced overload. We provide recursive equations and their solutions to calculate the steady-state fraction of surviving nodes, validate them through a set of simulation experiments, and discuss optimal load-capacity allocation strategies. Our results demonstrate that allocating the total excess capacity to each layer proportional to the mean effective load in the layer and distributing that excess capacity equally among the nodes within the layer ensures maximum robustness. The proposed framework for different failure conditions allows us to analyze the two overload conditions presented and can be extended to explore more complex interdependent relationships.

♻ ☆ Routing Guidance for Emerging Transportation Systems with Improved Dynamic Trip Equity

Ting Bai, Anni Li, Gehui Xu, Christos G. Cassandras, Andreas A. Malikopoulos

In this paper, we present a dynamic routing guidance system that optimizes route recommendations for individual vehicles within an emerging transportation system while enhancing travelers' trip equity. We develop a framework to quantify trip quality and equity in a dynamic travel environment, providing new insights into how routing guidance influences equity in road transportation. Our approach enables real-time routing by incorporating both monitored and anticipated traffic congestion. We provide conditions that ensure achieving perfect trip equity for all travelers in a free-flow network. Finally, simulation studies on 1,000 vehicles traversing an urban road network in Boston demonstrate that our proposed method improves trip equity by approximately 11.4\% compared to the shortest-route strategy. In addition, the results reveal that our approach redistributes travel costs across vehicle types through route optimization, contributing to a more equitable transportation system.

♻ ☆ Design, Implementation and Practical Energy-Efficiency Evaluation of a Blockchain Based Academic Credential Verification System for Low-Power Nodes

Gabriel Fernández-Blanco, Iván Froiz-Míguez, Paula Fraga-Lamas, Tiago M. Fernández-Caramés

The educational system manages extensive documentation and paperwork, which can lead to human errors and sometimes abuse or fraud, such as the falsification of diplomas, certificates or other credentials. In fact, in the last years, multiple cases of fraud have been detected, which have a significant cost to society, since they harm the trustworthiness of certificates and academic institutions. To tackle such an issue, this article proposes a solution aimed at recording and verifying academic records through a decentralized application that is supported by a smart contract deployed in the Ethereum blockchain and by a decentralized storage system based on Inter-Planetary File System (IPFS). The proposed solution is evaluated in terms of performance and energy-efficiency, comparing the results obtained with a traditional Proof-of-Work (PoW) consensus protocol and the new Proof-of-Authority (PoA) protocol. The results shown in this paper indicate that the latter is clearly greener and demands less CPU load. Moreover, this article compares the performance of a traditional computer and two SBCs (a Raspberry Pi 4 and an Orange Pi One), showing that is possible to make use of the latter low-power devices to implement blockchain nodes but at the cost of higher response latency. Furthermore, the impact of Ethereum gas limit is evaluated, demonstrating its significant influence on the blockchain network performance. Thus, this article provides guidelines, useful practical evaluations and key findings that will help the next generation of green blockchain developers and researchers.

♻ ☆ Gaussian Processes with Noisy Regression Inputs for Dynamical Systems

Tobias M. Wolff, Victor G. Lopez, Matthias A. Müller

This paper is centered around the approximation of dynamical systems by means of Gaussian processes. To this end, trajectories of such systems must be collected to be used as training data. The measurements of these trajectories are typically noisy, which implies that both the regression inputs and outputs are corrupted by noise. However, most of the literature considers only noise in the regression outputs. In this paper, we show how to account for the noise in the regression inputs in an extended Gaussian process framework to approximate scalar and multidimensional systems. We demonstrate the potential of our framework by comparing it to different state-of-the-art methods in several simulation examples.

comment: 6 pages

♻ ☆ Harmonic model predictive control for tracking sinusoidal references and its application to trajectory tracking

Pablo Krupa, Daniel Limon, Alberto Bemporad, Teodoro Alamo

Harmonic model predictive control (HMPC) is a recent model predictive control (MPC) formulation for tracking piece-wise constant references that includes a parameterized artificial harmonic reference as a decision variable, resulting in an increased performance and domain of attraction with respect to other MPC formulations. This article presents an extension of the HMPC formulation to track periodic harmonic/sinusoidal references and discusses its use for tracking arbitrary trajectories. The proposed formulation inherits the benefits of its predecessor, namely its good performance and large domain of attraction when using small prediction horizons, and that the complexity of its optimization problem does not depend on the period of the reference. We show closed-loop results discussing its performance and comparing it to other MPC formulations.

comment: Accepted version of the article published in IEEE Transactions on Automatic Control (8 pages, 5 figures)

♻ ☆ Data-Efficient Extremum-Seeking Control Using Kernel-Based Function Approximation

Wouter Weekers, Alessandro Saccon, Nathan van de Wouw

Existing extremum-seeking control (ESC) approaches typically rely on applying repeated perturbations to input parameters and performing measurements of the corresponding performance output. The required separation between the different timescales in the ESC loop means that performing these measurements can be time-consuming. Moreover, performing these measurements can be costly in practice, e.g., due to the use of resources. With these challenges in mind, it is desirable to reduce the number of measurements needed to optimize performance. Therefore, in this work, we present a sampled-data ESC approach aimed at reducing the number of measurements that need to be performed. In the proposed approach, we use input-output data obtained during regular operation of the extremum-seeking controller to construct online an approximation of the system's underlying cost function. By using this approximation to perform parameter updates when a decrease in the cost can be guaranteed, instead of performing additional measurements to perform this update, we make more efficient use of data collected during regular operation of the extremum-seeking controller. As a result, we indeed obtain a reduction in the required number of measurements to achieve optimization. We provide a stability analysis of the novel sampled-data ESC approach, and demonstrate the benefits of the synergy between kernel-based function approximation and standard ESC in simulation on a multi-input dynamical system.

comment: 16 pages, 5 figures, extended version of the paper submitted to Automatica

♻ ☆ Minimal positive Markov realizations

Hamed Taghavian, Jens Sjölund

Finding a positive state-space realization with the minimum dimension for a given transfer function is an open problem in control theory. In this paper, we focus on positive realizations in Markov form and propose a linear programming approach that computes them with a minimum dimension. Such minimum dimension of positive Markov realizations is an upper bound of the minimal positive realization dimension. However, we show that these two dimensions are equal for certain systems.

♻ ☆ Non-overshooting output shaping for switched linear systems under arbitrary switching using eigenstructure assignment

Kai Wulff, Maria Christine Honecker, Robert Schmid, Johann Reger

We consider the analytical control design for a pair of switched linear multiple-input multiple-output (MIMO) systems that are subject to arbitrary switching signals. A state feedback controller design method is proposed to obtain an eigenstructure assignment that ensures that the closed-loop switched system is globally asymptotically stable, and the outputs achieve the non-overshooting tracking of a step reference. Our analysis indicates whether non-overshooting or even monotonic tracking is achievable for the given system and considered outputs and provides a choice of possible eigenstructures to be assigned to the constituent subsystems. We derive a structural condition that verifies the feasibility of the chosen assignment. A constructive algorithm to obtain suitable feedback matrices is provided, and the method is illustrated with numerical examples.

♻ ☆ 1-2-3-Go! Policy Synthesis for Parameterized Markov Decision Processes via Decision-Tree Learning and Generalization

Muqsit Azeem, Debraj Chakraborty, Sudeep Kanav, Jan Kretinsky, Mohammadsadegh Mohagheghi, Stefanie Mohr, Maximilian Weininger

Despite the advances in probabilistic model checking, the scalability of the verification methods remains limited. In particular, the state space often becomes extremely large when instantiating parameterized Markov decision processes (MDPs) even with moderate values. Synthesizing policies for such \emph{huge} MDPs is beyond the reach of available tools. We propose a learning-based approach to obtain a reasonable policy for such huge MDPs. The idea is to generalize optimal policies obtained by model-checking small instances to larger ones using decision-tree learning. Consequently, our method bypasses the need for explicit state-space exploration of large models, providing a practical solution to the state-space explosion problem. We demonstrate the efficacy of our approach by performing extensive experimentation on the relevant models from the quantitative verification benchmark set. The experimental results indicate that our policies perform well, even when the size of the model is orders of magnitude beyond the reach of state-of-the-art analysis tools.

comment: Extended version of the paper accepted at VMCAI 2025

♻ ☆ Minimally Conservative Controlled-Invariant Set Synthesis Using Control Barrier Certificates

Naeim Ebrahimi Toulkani, Reza Ghabcheloo

Finding a controlled-invariant set for a system with state and control constraints is crucial for safety-critical applications. However, existing methods often produce overly conservative solutions. This paper presents a method for generating controlled-invariant (safe) sets for nonlinear polynomial control-affine systems using Control Barrier Certificates (CBCs). We formulate CBC conditions as Sum-of-Squares (SOS) constraints and solve them via an SOS Program (SOSP). First, we generalize existing SOSPs for CBC synthesis to handle environments with complex unsafe state representations. Then, we propose an iterative algorithm that progressively enlarges the safe set constructed by the synthesized CBCs by maximizing boundary expansion at each iteration. We theoretically prove that our method guarantees strict safe set expansion at every step. Finally, we validate our approach with numerical simulations in 2D and 3D for single-input and multi-input systems. Empirical results show that the safe set generated by our method covers in most part a larger portion of the state space compared to two state-of-the-art techniques.

♻ ☆ Designing Heterogeneous GNNs with Desired Permutation Properties for Wireless Resource Allocation

Jianyu Zhao, Chenyang Yang, Tingting Liu

Graph neural networks (GNNs) have been designed for learning a variety of wireless policies, i.e., the mappings from environment parameters to decision variables, thanks to their superior performance, and the potential in enabling scalability and size generalizability. These merits are rooted in leveraging permutation prior, i.e., satisfying the permutation property of the policy to be learned (referred to as desired permutation property). Many wireless policies are with complicated permutation properties. To satisfy these properties, heterogeneous GNNs (HetGNNs) should be used to learn such policies. There are two critical factors that enable a HetGNN to satisfy a desired permutation property: constructing an appropriate heterogeneous graph and judiciously designing the architecture of the HetGNN. However, both the graph and the HetGNN are designed heuristically so far. In this paper, we strive to provide a systematic approach for the design to satisfy the desired permutation property. We first propose a method for constructing a graph for a policy, where the edges and their types are defined for the sake of satisfying complicated permutation properties. Then, we provide and prove three sufficient conditions to design a HetGNN such that it can satisfy the desired permutation property when learning over an appropriate graph. These conditions suggest a method of designing the HetGNN with desired permutation property by sharing the processing, combining, and pooling functions according to the types of vertices and edges of the graph. We take power allocation and hybrid precoding policies as examples for demonstrating how to apply the proposed methods and validating the impact of the permutation prior by simulations.

♻ ☆ Markov $α$-Potential Games

Xin Guo, Xinyu Li, Chinmay Maheshwari, Shankar Sastry, Manxi Wu

We propose a new framework of Markov $\alpha$-potential games to study Markov games. We show that any Markov game with finite-state and finite-action is a Markov $\alpha$-potential game, and establish the existence of an associated $\alpha$-potential function. Any optimizer of an $\alpha$-potential function is shown to be an $\alpha$-stationary Nash equilibrium. We study two important classes of practically significant Markov games, Markov congestion games and the perturbed Markov team games, via the framework of Markov $\alpha$-potential games, with explicit characterization of an upper bound for $\alpha$ and its relation to game parameters. Additionally, we provide a semi-infinite linear programming based formulation to obtain an upper bound for $\alpha$ for any Markov game. Furthermore, we study two equilibrium approximation algorithms, namely the projected gradient-ascent algorithm and the sequential maximum improvement algorithm, along with their Nash regret analysis, and corroborate the results with numerical experiments.

comment: 33 pages, 5 figures

♻ ☆ Decentralized State Estimation and Opacity Verification Based on Partially Ordered Observation Sequences

Dajiang Sun, Christoforos N. Hadjicostis, Zhiwu Li

In this paper, we investigate state estimation and opacity verification problems within a decentralized observation architecture. Specifically, we consider a discrete event system whose behavior is recorded by a set of observation sites. These sites transmit the partially ordered sequences of observations that they record to a coordinator whenever a \textit{synchronization} occurs. To properly analyze the system behavior from the coordinator's viewpoint, we first introduce the notion of a Complete Synchronizing Sequence structure (CSS structure), which concisely captures the state evolution of each system state upon different information provided by the observation sites. Based on the CSS structure, we then construct corresponding current-state and initial-state estimators for offline state estimation at the coordinator. When used to verify state-isolation properties under this decentralized architecture, the use of CSS structure demonstrates a significant reduction in complexity compared with existing approaches in the literature. In particular, we discuss how to verify initial-state opacity at the coordinator, as well as a novel opacity notion, namely current-state-at-synchronization opacity.

♻ ☆ Representation and Stability Analysis of 1D PDEs with Periodic Boundary Conditions

Declan Jagt, Sergei Chernyshenko, Matthew Peet

PDEs with periodic boundary conditions are frequently used to model processes in large spatial environments, assuming solutions to extend periodically beyond some bounded interval. However, for 2nd order PDEs with periodic boundary conditions, the nullspace of the differential operator $\frac{\partial^2}{\partial x^2}$ is nontrivial on the PDE domain, and solutions may converge to non-stationary trajectories existing in this nullspace. To test this convergence behaviour, in this paper, it is shown how we can model these trajectories for a broad class of linear, 2nd order, 1D PDEs with periodic as well as more general boundary conditions, using the Partial Integral Equation (PIE) representation. In particular, it is first shown how any function $\mathbf{u}(t)$ in the PDE domain can be decomposed into a component defined by $\mathbf{u}_{xx}(t)$, and a component $\bar{\mathbf{u}}(t)$ existing in the nullspace of $\frac{\partial^2}{\partial x^2}$. An equivalent representation of linear PDEs is then derived as a PIE, explicitly defining the dynamics of both $\mathbf{u}_{xx}(t)$ and $\bar{\mathbf{u}}(t)$. Finally, a notion of exponential stability is defined for trajectories $\mathbf{u}^*(t)=\bar{\mathbf{u}}(t)$, and it is shown that stability of these trajectories as well as of the equilibrium $\mathbf{u}^*\equiv0$ can be tested by solving a linear operator inequality. The proposed methodology is applied to two examples, demonstrating that stability can be verified with tight bounds on the rate of exponential decay.

♻ ☆ Battery Operations in Electricity Markets: Strategic Behavior and Distortions

Jerry Anunrojwong, Santiago R. Balseiro, Omar Besbes, Bolun Xu

Electric power systems are undergoing a major transformation as they integrate intermittent renewable energy sources, and batteries to smooth out variations in renewable energy production. As privately-owned batteries grow from their role as marginal "price-takers" to significant players in the market, a natural question arises: How do batteries operate in electricity markets, and how does the strategic behavior of decentralized batteries distort decisions compared to centralized batteries? We propose an analytically tractable model that captures salient features of the highly complex electricity market. We derive in closed form the resulting battery behavior and generation cost in three operating regimes: (i) no battery, (ii) centralized battery, and (ii) decentralized profit-maximizing battery. We establish that a decentralized battery distorts its discharge decisions in three ways. First, there is quantity withholding, i.e., discharging less than centrally optimal. Second, there is a shift in participation from day-ahead to real-time, i.e., postponing some of its discharge from day-ahead to real-time. Third, there is reduction in real-time responsiveness, or discharging less in response to smoothing real-time demand than centrally optimal. We also quantify the impact of the battery market power on total system cost via the Price of Anarchy metric, and prove that the it is always between $9/8$ and $4/3$. That is, incentive misalignment always exists, but it is bounded even in the worst case. We calibrate our model to real data from Los Angeles and Houston. Lastly, we show that competition is very effective at reducing distortions, but many market power mitigation mechanisms backfire, and lead to higher total cost.

♻ ☆ Independent and Decentralized Learning in Markov Potential Games

Chinmay Maheshwari, Manxi Wu, Druv Pai, Shankar Sastry

We study a multi-agent reinforcement learning dynamics, and analyze its asymptotic behavior in infinite-horizon discounted Markov potential games. We focus on the independent and decentralized setting, where players do not know the game parameters, and cannot communicate or coordinate. In each stage, players update their estimate of Q-function that evaluates their total contingent payoff based on the realized one-stage reward in an asynchronous manner. Then, players independently update their policies by incorporating an optimal one-stage deviation strategy based on the estimated Q-function. Inspired by the actor-critic algorithm in single-agent reinforcement learning, a key feature of our learning dynamics is that agents update their Q-function estimates at a faster timescale than the policies. Leveraging tools from two-timescale asynchronous stochastic approximation theory, we characterize the convergent set of learning dynamics.

comment: 43 pages, 1 figure

♻ ☆ Convergence of Decentralized Actor-Critic Algorithm in General-sum Markov Games

Chinmay Maheshwari, Manxi Wu, Shankar Sastry

Markov games provide a powerful framework for modeling strategic multi-agent interactions in dynamic environments. Traditionally, convergence properties of decentralized learning algorithms in these settings have been established only for special cases, such as Markov zero-sum and potential games, which do not fully capture real-world interactions. In this paper, we address this gap by studying the asymptotic properties of learning algorithms in general-sum Markov games. In particular, we focus on a decentralized algorithm where each agent adopts an actor-critic learning dynamic with asynchronous step sizes. This decentralized approach enables agents to operate independently, without requiring knowledge of others' strategies or payoffs. We introduce the concept of a Markov Near-Potential Function (MNPF) and demonstrate that it serves as an approximate Lyapunov function for the policy updates in the decentralized learning dynamics, which allows us to characterize the convergent set of strategies. We further strengthen our result under specific regularity conditions and with finite Nash equilibria.

comment: 18 pages, 3 figure

♻ ☆ Low-Rank Thinning

Annabelle Michael Carrell, Albert Gong, Abhishek Shetty, Raaz Dwivedi, Lester Mackey

The goal in thinning is to summarize a dataset using a small set of representative points. Remarkably, sub-Gaussian thinning algorithms like Kernel Halving and Compress can match the quality of uniform subsampling while substantially reducing the number of summary points. However, existing guarantees cover only a restricted range of distributions and kernel-based quality measures and suffer from pessimistic dimension dependence. To address these deficiencies, we introduce a new low-rank analysis of sub-Gaussian thinning that applies to any distribution and any kernel, guaranteeing high-quality compression whenever the kernel or data matrix is approximately low-rank. To demonstrate the broad applicability of the techniques, we design practical sub-Gaussian thinning approaches that improve upon the best known guarantees for approximating attention in transformers, accelerating stochastic gradient training through reordering, and distinguishing distributions in near-linear time.

♻ ☆ An Optimistic-Robust Approach for Dynamic Positioning of Omnichannel Inventories

Pavithra Harsha, Shivaram Subramanian, Ali Koc, Mahesh Ramakrishna, Brian Quanz, Dhruv Shah, Chandra Narayanaswami

We introduce a new class of data-driven and distribution-free optimistic-robust bimodal inventory optimization (BIO) strategy to effectively allocate inventory across a retail chain to meet time-varying, uncertain omnichannel demand. The bimodal nature of BIO stems from its ability to balance downside risk, as in traditional Robust Optimization (RO), which focuses on worst-case adversarial demand, with upside potential to enhance average-case performance. This enables BIO to remain as resilient as RO while capturing benefits that would otherwise be lost due to endogenous outliers. Omnichannel inventory planning provides a suitable problem setting for analyzing the effectiveness of BIO's bimodal strategy in managing the tradeoff between lost sales at stores and cross-channel e-commerce fulfillment costs, factors that are inherently asymmetric due to channel-specific behaviors. We provide structural insights about the BIO solution and how it can be tuned to achieve a preferred tradeoff between robustness and the average-case performance. Using a real-world dataset from a large American omnichannel retail chain, a business value assessment during a peak period indicates that BIO outperforms pure RO by 27% in terms of realized average profitability and surpasses other competitive baselines under imperfect distributional information by over 10%. This demonstrates that BIO provides a novel, data-driven, and distribution-free alternative to traditional RO that achieves strong average performance while carefully balancing robustness.

♻ ☆ Optimizing Posterior Samples for Bayesian Optimization via Rootfinding ICLR 2025

Taiwo A. Adebiyi, Bach Do, Ruda Zhang

Bayesian optimization devolves the global optimization of a costly objective function to the global optimization of a sequence of acquisition functions. This inner-loop optimization can be catastrophically difficult if it involves posterior sample paths, especially in higher dimensions. We introduce an efficient global optimization strategy for posterior samples based on global rootfinding. It provides gradient-based optimizers with two sets of judiciously selected starting points, designed to combine exploration and exploitation. The number of starting points can be kept small without sacrificing optimization quality. Remarkably, even with just one point from each set, the global optimum is discovered most of the time. The algorithm scales practically linearly to high dimensions, breaking the curse of dimensionality. For Gaussian process Thompson sampling (GP-TS), we demonstrate remarkable improvement in both inner- and outer-loop optimization, surprisingly outperforming alternatives like EI and GP-UCB in most cases. Our approach also improves the performance of other posterior sample-based acquisition functions, such as variants of entropy search. Furthermore, we propose a sample-average formulation of GP-TS, which has a parameter to explicitly control exploitation and can be computed at the cost of one posterior sample. Our implementation is available at https://github.com/UQUH/TSRoots .

comment: Published at the Thirteenth International Conference on Learning Representations (ICLR 2025)

♻ ☆ A general framework for inexact splitting algorithms with relative errors and applications to Chambolle-Pock and Davis-Yin methods

M. Marques Alves, Dirk A. Lorenz, Emanuele Naldi

In this work we apply the recently introduced framework of degenerate preconditioned proximal point algorithms to the hybrid proximal extragradient (HPE) method for maximal monotone inclusions. The latter is a method that allows inexact proximal (or resolvent) steps where the error is controlled by a relative-error criterion. Recently the HPE framework has been extended to the Douglas-Rachford method by Eckstein and Yao. In this paper we further extend the applicability of the HPE framework to splitting methods. To this end we use the framework of degenerate preconditioners that allows to write a large class of splitting methods as preconditioned proximal point algorithms. In this way, we modify many splitting methods such that one or more of the resolvents can be computed inexactly with an error that is controlled by an adaptive criterion. Further, we illustrate the algorithmic framework in the case of Chambolle-Pock's primal dual hybrid gradient method and the Davis-Yin's forward Douglas-Rachford method. In both cases, the inexact computation of the resolvent shows clear advantages in computing time and accuracy.

comment: 34 pages, 6 figures

♻ ☆ Cislunar Space Situational Awareness Constellation Design and Planning with Facility Location Problem

Yuri Shimane, Kento Tomita, Koki Ho

Driven by the surmounting interest for dedicated infrastructure in cislunar space, this work considers the satellite constellation design for cislunar space situational awareness (CSSA). We propose a mixed-integer linear programming (MILP)-based formulation that simultaneously tackles the constellation design and sensor-tasking subproblems surrounding CSSA. Our approach generates constellation designs that provide coverage with considerations for the field-of-view of observers. We propose a time-expanded p-Median problem (TE-p-MP) which considers the optimal placement of p space-based observers into discretized locations based on orbital slots along libration point orbits, simultaneously with observer pointing directions across discretized time. We further develop a Lagrangian method for the TE-p-MP, where a relaxed problem with an analytical solution is derived, and customized heuristics leveraging the orbital structure of candidate observer locations are devised. The performance of the proposed formulation is demonstrated with several case studies for CSSA constellations monitoring the cislunar Cone of Shame and a periodic time-varying transit window for low-energy transfers located in the Earth-Moon L2 neck region. The proposed problem formulation, along with the Lagrangian method, is demonstrated to enable a fast assessment of near-optimal CSSA constellations, equipping decision-makers with a critical technique for exploring the design trade space.

comment: 50 pages, 15 figures

♻ ☆ Quantum algorithms for optimizers

Giacomo Nannicini

This is a set of lecture notes for a Ph.D.-level course on quantum algorithms, with an emphasis on quantum optimization algorithms. It is developed for applied mathematicians and engineers, and requires no previous background in quantum mechanics. The main topics of this course, in addition to a rigorous introduction to the computational model, are: input/output models, quantum search, the quantum gradient algorithm, matrix manipulation algorithms, the mirror descent framework for semidefinite optimization (including the matrix multiplicative weights update algorithm), adiabatic optimization.

comment: Fixed several typos, rewrote chapter on SDP to include mirror descent framework

♻ ☆ Gradient Descent for Convex and Smooth Noisy Optimization

Feifei Hu, Mathieu Gerber

We study the use of gradient descent with backtracking line search (GD-BLS) to solve the noisy optimization problem $\theta_\star:=\mathrm{argmin}_{\theta\in\mathbb{R}^d} \mathbb{E}[f(\theta,Z)]$, imposing that the function $F(\theta):=\mathbb{E}[f(\theta,Z)]$ is strictly convex but not necessarily $L$-smooth. Assuming that $\mathbb{E}[\|\nabla_\theta f(\theta_\star,Z)\|^2]<\infty$, we first prove that sample average approximation based on GD-BLS allows to estimate $\theta_\star$ with an error of size $\mathcal{O}_{\mathbb{P}}(B^{-0.25})$, where $B$ is the available computational budget. We then show that we can improve upon this rate by stopping the optimization process earlier when the gradient of the objective function is sufficiently close to zero, and use the residual computational budget to optimize, again with GD-BLS, a finer approximation of $F$. By iteratively applying this strategy $J$ times, we establish that we can estimate $\theta_\star$ with an error of size $\mathcal{O}_{\mathbb{P}}(B^{-\frac{1}{2}(1-\delta^{J})})$, where $\delta\in(1/2,1)$ is a user-specified parameter. More generally, we show that if $\mathbb{E}[\|\nabla_\theta f(\theta_\star,Z)\|^{1+\alpha}]<\infty$ for some known $\alpha\in (0,1]$ then this approach, which can be seen as a retrospective approximation algorithm with a fixed computational budget, allows to learn $\theta_\star$ with an error of size $\mathcal{O}_{\mathbb{P}}(B^{-\frac{\alpha}{1+\alpha}(1-\delta^{J})})$, where $\delta\in(2\alpha/(1+3\alpha),1)$ is a tuning parameter. Beyond knowing $\alpha$, achieving the aforementioned convergence rates do not require to tune the algorithms parameters according to the specific functions $F$ and $f$ at hand, and we exhibit a simple noisy optimization problem for which stochastic gradient is not guaranteed to converge while the algorithms discussed in this work are.

comment: 40 pages, 3 figures

♻ ☆ One Benders cut to rule all schedules in the neighbourhood

Ioannis Avgerinos, Ioannis Mourtos, Stavros Vatikiotis, Georgios Zois

Logic-Based Benders Decomposition (LBBD) and its Branch-and-Cut variant, namely Branch-and-Check, enjoy an extensive applicability on a broad variety of problems, including scheduling. Although LBBD offers problem-specific cuts to impose tighter dual bounds, its application to resource-constrained scheduling remains less explored. Given a position-based Mixed-Integer Linear Programming (MILP) formulation for scheduling on unrelated parallel machines, we notice that certain $k-$OPT neighbourhoods could implicitly be explored by regular local search operators, thus allowing us to integrate Local Branching into Branch-and-Check schemes. After enumerating such neighbourhoods and obtaining their local optima - hence, proving that they are suboptimal - a local branching cut (applied as a Benders cut) eliminates all their solutions at once, thus avoiding an overload of the master problem with thousands of Benders cuts. However, to guarantee convergence to optimality, the constructed neighbourhood should be exhaustively explored, hence this time-consuming procedure must be accelerated by domination rules or selectively implemented on nodes which are more likely to reduce the optimality gap. In this study, the realisation of this idea is limited on the common 'internal (job) swaps' to construct formulation-specific $4$-OPT neighbourhoods. Nonetheless, the experimentation on two challenging scheduling problems (i.e., the minimisation of total completion times and the minimisation of total tardiness on unrelated machines with sequence-dependent and resource-constrained setups) shows that the proposed methodology offers considerable reductions of optimality gaps or faster convergence to optimality. The simplicity of our approach allows its transferability to other neighbourhoods and different sequencing optimisation problems, hence providing a promising prospect to improve Branch-and-Check methods.

♻ ☆ Optimization Insights into Deep Diagonal Linear Networks

Hippolyte Labarrière, Cesare Molinari, Lorenzo Rosasco, Silvia Villa, Cristian Vega

Overparameterized models trained with (stochastic) gradient descent are ubiquitous in modern machine learning. These large models achieve unprecedented performance on test data, but their theoretical understanding is still limited. In this paper, we take a step towards filling this gap by adopting an optimization perspective. More precisely, we study the implicit regularization properties of the gradient flow "algorithm" for estimating the parameters of a deep diagonal neural network. Our main contribution is showing that this gradient flow induces a mirror flow dynamic on the model, meaning that it is biased towards a specific solution of the problem depending on the initialization of the network. Along the way, we prove several properties of the trajectory.

♻ ☆ A stochastic gradient descent algorithm with random search directions

Eméric Gbaguidi

Stochastic coordinate descent algorithms are efficient methods in which each iterate is obtained by fixing most coordinates at their values from the current iteration, and approximately minimizing the objective with respect to the remaining coordinates. However, this approach is usually restricted to canonical basis vectors of $\mathbb{R}^d$. In this paper, we develop a new class of stochastic gradient descent algorithms with random search directions which uses the directional derivative of the gradient estimate following more general random vectors. We establish the almost sure convergence of these algorithms with decreasing step. We further investigate their central limit theorem and pay particular attention to analyze the impact of the search distributions on the asymptotic covariance matrix. We also provide non-asymptotic $\mathbb{L}^p$ rates of convergence.

♻ ☆ Turnpike and dissipativity in generalized discrete-time stochastic linear-quadratic optimal control

Jonas Schießl, Ruchuan Ou, Timm Faulwasser, Michael Heinrich Baumann, Lars Grüne

We investigate different turnpike phenomena of generalized discrete-time stochastic linear-quadratic optimal control problems. Our analysis is based on a novel strict dissipativity notion for such problems, in which a stationary stochastic process replaces the optimal steady state of the deterministic setting. We show that from this time-varying dissipativity notion, we can conclude turnpike behaviors concerning different objects like distributions, moments, or sample paths of the stochastic system and that the distributions of the stationary pair can be characterized by a stationary optimization problem. The analytical findings are illustrated by numerical simulations.

♻ ☆ Projection-based Reduced Order Modelling for Unsteady Parametrized Optimal Control Problems in 3D Cardiovascular Flows

Surabhi Rathore, Pasquale Claudio Africa, Francesco Ballarin, Federico Pichi, Michele Girfoglio, Gianluigi Rozza

This paper presents a projection-based reduced order modelling (ROM) framework for unsteady parametrized optimal control problems (OCP$_{(\mu)}$s) arising from cardiovascular (CV) applications. In real-life scenarios, accurately defining outflow boundary conditions in patient-specific models poses significant challenges due to complex vascular morphologies, physiological conditions, and high computational demands. These challenges make it difficult to compute realistic and reliable CV hemodynamics by incorporating clinical data such as 4D magnetic resonance imaging. To address these challenges, we focus on controlling the outflow boundary conditions to optimize CV flow dynamics and minimize the discrepancy between target and computed flow velocity profiles. The fluid flow is governed by unsteady Navier--Stokes equations with physical parametric dependence, i.e. the Reynolds number. Numerical solutions of OCP$_{(\mu)}$s require substantial computational resources, highlighting the need for robust and efficient ROMs to perform real-time and many-query simulations. Here, we aim at investigating the performance of a projection-based reduction technique that relies on the offline-online paradigm, enabling significant computational cost savings. The Galerkin finite element method is used to compute the high-fidelity solutions in the offline phase. We implemented a nested-proper orthogonal decomposition (nested-POD) for fast simulation of OCP$_{(\mu)}$s that encompasses two stages: temporal compression for reducing dimensionality in time, followed by parametric-space compression on the precomputed POD modes. We tested the efficacy of the methodology on vascular models, namely an idealized bifurcation geometry and a patient-specific coronary artery bypass graft, incorporating stress control at the outflow boundary, observing consistent speed-up with respect to high-fidelity strategies.

♻ ☆ Joint Replenishment Strategy for Multiple Satellite Constellations with Shared Launch Opportunities

Jaewoo Kim, Taehyun Sung, Woonam Hwang, Jaemyung Ahn

This paper proposes a novel replenishment strategy that can jointly support multiple satellite constellations. In this approach, multiple constellations share launch opportunities and parking orbits to address the operational satellite failures and ensure the desired service level of the constellations. We develop an inventory management model based on parametric replenishment policies, considering the launch vehicle's capacity and the shipping size of satellites. Based on this model, we introduce two decision-making scenarios and propose their corresponding solution frameworks. We conduct two case studies to provide valuable insights into the proposed strategy and demonstrate its applicability to supply chain management for maintaining multiple satellite constellations.

comment: 36 pages, 6 figures

♻ ☆ Multi-Agent Relative Investment Games in a Jump Diffusion Market with Deep Reinforcement Learning Algorithm

Liwei Lu, Ruimeng Hu, Xu Yang, Yi Zhu

This paper focuses on multi-agent stochastic differential games for jump-diffusion systems. On one hand, we study the multi-agent game for optimal investment in a jump-diffusion market. We derive constant Nash equilibria and provide sufficient conditions for their existence and uniqueness for exponential, power, and logarithmic utilities, respectively. On the other hand, we introduce a computational framework based on the actor-critic method in deep reinforcement learning to solve the stochastic control problem with jumps. We extend this algorithm to address the multi-agent game with jumps and utilize parallel computing to enhance computational efficiency. We present numerical examples of the Merton problem with jumps, linear quadratic regulators, and the optimal investment game under various settings to demonstrate the accuracy, efficiency, and robustness of the proposed method. In particular, neural network solutions numerically converge to the derived constant Nash equilibrium for the multi-agent game.

♻ ☆ Non-overshooting output shaping for switched linear systems under arbitrary switching using eigenstructure assignment

Kai Wulff, Maria Christine Honecker, Robert Schmid, Johann Reger

We consider the analytical control design for a pair of switched linear multiple-input multiple-output (MIMO) systems that are subject to arbitrary switching signals. A state feedback controller design method is proposed to obtain an eigenstructure assignment that ensures that the closed-loop switched system is globally asymptotically stable, and the outputs achieve the non-overshooting tracking of a step reference. Our analysis indicates whether non-overshooting or even monotonic tracking is achievable for the given system and considered outputs and provides a choice of possible eigenstructures to be assigned to the constituent subsystems. We derive a structural condition that verifies the feasibility of the chosen assignment. A constructive algorithm to obtain suitable feedback matrices is provided, and the method is illustrated with numerical examples.

♻ ☆ $p$-Adic Polynomial Regression as Alternative to Neural Network for Approximating $p$-Adic Functions of Many Variables

Alexander P. Zubarev

A method for approximating continuous functions $\mathbb{Z}_{p}^{n}\rightarrow\mathbb{Z}_{p}$ by a linear superposition of continuous functions $\mathbb{Z}_{p}\rightarrow\mathbb{Z}_{p}$ is presented and a polynomial regression model is constructed that allows approximating such functions with any degree of accuracy. A physical interpretation of such a model is given and possible methods for its training are discussed. The proposed model can be considered as a simple alternative to possible $p$-adic models based on neural network architecture.

comment: 10 pages

♻ ☆ An $α$-potential game framework for $N$-player dynamic games

Xin Guo, Xinyu Li, Yufei Zhang

This paper proposes and studies a general form of dynamic $N$-player non-cooperative games called $\alpha$-potential games, where the change of a player's value function upon her unilateral deviation from her strategy is equal to the change of an $\alpha$-potential function up to an error $\alpha$. Analogous to the static potential game (which corresponds to $\alpha=0$), the $\alpha$-potential game framework is shown to reduce the challenging task of finding $\alpha$-Nash equilibria for a dynamic game to minimizing the $\alpha$-potential function. Moreover, an analytical characterization of $\alpha$-potential functions is established, with $\alpha$ represented in terms of the magnitude of the asymmetry of value functions' second-order derivatives. For stochastic differential games in which the state dynamic is a controlled diffusion, $\alpha$ is characterized in terms of the number of players, the choice of admissible strategies, and the intensity of interactions and the level of heterogeneity among players. Two classes of stochastic differential games, namely distributed games and games with mean field interactions, are analyzed to highlight the dependence of $\alpha$ on general game characteristics that are beyond the mean-field paradigm, which focuses on the limit of $N$ with homogeneous players. To analyze the $\alpha$-NE, the associated optimization problem is embedded into a conditional McKean-Vlasov control problem. A verification theorem is established to construct $\alpha$-NE based on solutions to an infinite-dimensional Hamilton-Jacobi-Bellman equation, which is reduced to a system of ordinary differential equations for linear-quadratic games.

comment: 36 pages

♻ ☆ A point cloud reconstruction method based on uncertainty feature enhancement for aerodynamic shape optimization

Junlin Li, Yang Zhang, Bo Pang, Junqiang Bai, Jiakuan Xu

The precision of shape representation and the dimensionality of the design space significantly influence the cost and outcomes of aerodynamic optimization. The design space can be represented more compactly by maintaining geometric precision while reducing dimensions, hence enhancing the cost-effectiveness of the optimization process. This research presents a new point cloud autoencoder architecture, called AE-BUFE, designed to attain efficient and precise generalized representations of 3D aircraft through uncertainty analysis of the deformation relationships among surface grid points. The deep learning architecture consists of two components: the uncertainty index-based feature enhancement module and the point cloud autoencoder module. It learns the shape features of the point cloud geometric representation to establish a low-dimensional latent space. To assess and evaluate the efficiency of the method, a comparison was conducted with the prevailing point cloud autoencoder architecture and the proper orthogonal decomposition (POD) linear dimensionality reduction method under conditions of complex shape deformation. The results showed that the new architecture significantly improved the extraction effect of the low-dimensional latent space. Then, we developed the SBO optimization framework based on the AE-BUFE parameterization method and completed a multi-objective aerodynamic optimization design for a wide-speed-range vehicle considering volume and moment constraints. While ensuring the take-off and landing performance, the aerodynamic performance is improved at transonic and hypersonic conditions, which verifies the efficiency and engineering practicability of this method.

comment: The data has some error which can lead to wrong understanding

♻ ☆ Representation and Stability Analysis of 1D PDEs with Periodic Boundary Conditions

Declan Jagt, Sergei Chernyshenko, Matthew Peet

PDEs with periodic boundary conditions are frequently used to model processes in large spatial environments, assuming solutions to extend periodically beyond some bounded interval. However, for 2nd order PDEs with periodic boundary conditions, the nullspace of the differential operator $\frac{\partial^2}{\partial x^2}$ is nontrivial on the PDE domain, and solutions may converge to non-stationary trajectories existing in this nullspace. To test this convergence behaviour, in this paper, it is shown how we can model these trajectories for a broad class of linear, 2nd order, 1D PDEs with periodic as well as more general boundary conditions, using the Partial Integral Equation (PIE) representation. In particular, it is first shown how any function $\mathbf{u}(t)$ in the PDE domain can be decomposed into a component defined by $\mathbf{u}_{xx}(t)$, and a component $\bar{\mathbf{u}}(t)$ existing in the nullspace of $\frac{\partial^2}{\partial x^2}$. An equivalent representation of linear PDEs is then derived as a PIE, explicitly defining the dynamics of both $\mathbf{u}_{xx}(t)$ and $\bar{\mathbf{u}}(t)$. Finally, a notion of exponential stability is defined for trajectories $\mathbf{u}^*(t)=\bar{\mathbf{u}}(t)$, and it is shown that stability of these trajectories as well as of the equilibrium $\mathbf{u}^*\equiv0$ can be tested by solving a linear operator inequality. The proposed methodology is applied to two examples, demonstrating that stability can be verified with tight bounds on the rate of exponential decay.

♻ ☆ A novel algorithm for optimizing bundle adjustment in image sequence alignment

Hailin Xu, Hongxia Wang, Huanshui Zhang

The Bundle Adjustment (BA) model is commonly optimized using a nonlinear least squares method, with the Levenberg-Marquardt (L-M) algorithm being a typical choice. However, despite the L-M algorithm's effectiveness, its sensitivity to initial conditions often results in slower convergence when applied to poorly conditioned datasets, motivating the exploration of alternative optimization strategies. This paper introduces a novel algorithm for optimizing the BA model in the context of image sequence alignment for cryo-electron tomography, utilizing optimal control theory to directly optimize general nonlinear functions. The proposed Optimal Control Algorithm (OCA) exhibits superior convergence rates and effectively mitigates the oscillatory behavior frequently observed in L-M algorithm. Extensive experiments on both synthetic and real-world datasets were conducted to evaluate the algorithm's performance. The results demonstrate that the OCA achieves faster convergence compared to the L-M algorithm. Moreover, the incorporation of a bisection-based update procedure significantly enhances the OCA's performance, particularly in poorly initialized datasets. These findings indicate that the OCA can substantially improve the efficiency of 3D reconstructions in cryo-electron tomography.

♻ ☆ Convergence of Decentralized Actor-Critic Algorithm in General-sum Markov Games

Chinmay Maheshwari, Manxi Wu, Shankar Sastry

Markov games provide a powerful framework for modeling strategic multi-agent interactions in dynamic environments. Traditionally, convergence properties of decentralized learning algorithms in these settings have been established only for special cases, such as Markov zero-sum and potential games, which do not fully capture real-world interactions. In this paper, we address this gap by studying the asymptotic properties of learning algorithms in general-sum Markov games. In particular, we focus on a decentralized algorithm where each agent adopts an actor-critic learning dynamic with asynchronous step sizes. This decentralized approach enables agents to operate independently, without requiring knowledge of others' strategies or payoffs. We introduce the concept of a Markov Near-Potential Function (MNPF) and demonstrate that it serves as an approximate Lyapunov function for the policy updates in the decentralized learning dynamics, which allows us to characterize the convergent set of strategies. We further strengthen our result under specific regularity conditions and with finite Nash equilibria.

comment: 18 pages, 3 figure

♻ ☆ ActiveGAMER: Active GAussian Mapping through Efficient Rendering CVPR2025

Liyan Chen, Huangying Zhan, Kevin Chen, Xiangyu Xu, Qingan Yan, Changjiang Cai, Yi Xu

We introduce ActiveGAMER, an active mapping system that utilizes 3D Gaussian Splatting (3DGS) to achieve high-quality, real-time scene mapping and exploration. Unlike traditional NeRF-based methods, which are computationally demanding and restrict active mapping performance, our approach leverages the efficient rendering capabilities of 3DGS, allowing effective and efficient exploration in complex environments. The core of our system is a rendering-based information gain module that dynamically identifies the most informative viewpoints for next-best-view planning, enhancing both geometric and photometric reconstruction accuracy. ActiveGAMER also integrates a carefully balanced framework, combining coarse-to-fine exploration, post-refinement, and a global-local keyframe selection strategy to maximize reconstruction completeness and fidelity. Our system autonomously explores and reconstructs environments with state-of-the-art geometric and photometric accuracy and completeness, significantly surpassing existing approaches in both aspects. Extensive evaluations on benchmark datasets such as Replica and MP3D highlight ActiveGAMER's effectiveness in active mapping tasks.

comment: Accepted to CVPR2025

♻ ☆ NTSEBENCH: Cognitive Reasoning Benchmark for Vision Language Models

Pranshu Pandya, Vatsal Gupta, Agney S Talwarr, Tushar Kataria, Dan Roth, Vivek Gupta

Cognitive textual and visual reasoning tasks, including puzzles, series, and analogies, demand the ability to quickly reason, decipher, and evaluate patterns both textually and spatially. Due to extensive training on vast amounts of human-curated data, LLMs and VLMs excel in common-sense reasoning tasks, however still struggle with more complex reasoning that demands deeper cognitive understanding. We introduce NTSEBench, a new dataset designed to evaluate cognitive multi-modal reasoning and problem-solving skills of large models. The dataset contains 2728 multiple-choice questions, accompanied by a total of 4,642 images, categorized into 26 different types. These questions are drawn from the nationwide NTSE examination in India and feature a mix of visual and textual general aptitude challenges, designed to assess intelligence and critical thinking skills beyond mere rote learning. We establish baselines on the dataset using state-of-the-art LLMs and VLMs. To facilitate a comparison between open source and propriety models, we propose four distinct modeling strategies to handle different modalities -- text and images -- in the dataset instances.

comment: 28 pages, 3 figures, 12 tables

♻ ☆ Rehearsal-free Federated Domain-incremental Learning

Rui Sun, Haoran Duan, Jiahua Dong, Varun Ojha, Tejal Shah, Rajiv Ranjan

We introduce a rehearsal-free federated domain incremental learning framework, RefFiL, based on a global prompt-sharing paradigm to alleviate catastrophic forgetting challenges in federated domain-incremental learning, where unseen domains are continually learned. Typical methods for mitigating forgetting, such as the use of additional datasets and the retention of private data from earlier tasks, are not viable in federated learning (FL) due to devices' limited resources. Our method, RefFiL, addresses this by learning domain-invariant knowledge and incorporating various domain-specific prompts from the domains represented by different FL participants. A key feature of RefFiL is the generation of local fine-grained prompts by our domain adaptive prompt generator, which effectively learns from local domain knowledge while maintaining distinctive boundaries on a global scale. We also introduce a domain-specific prompt contrastive learning loss that differentiates between locally generated prompts and those from other domains, enhancing RefFiL's precision and effectiveness. Compared to existing methods, RefFiL significantly alleviates catastrophic forgetting without requiring extra memory space, making it ideal for privacy-sensitive and resource-constrained devices.

comment: Camera ready version. Accepted by the IEEE ICDCS, 2025

♻ ☆ DetailGen3D: Generative 3D Geometry Enhancement via Data-Dependent Flow

Ken Deng, Yuan-Chen Guo, Jingxiang Sun, Zi-Xin Zou, Yangguang Li, Xin Cai, Yan-Pei Cao, Yebin Liu, Ding Liang

Modern 3D generation methods can rapidly create shapes from sparse or single views, but their outputs often lack geometric detail due to computational constraints. We present DetailGen3D, a generative approach specifically designed to enhance these generated 3D shapes. Our key insight is to model the coarse-to-fine transformation directly through data-dependent flows in latent space, avoiding the computational overhead of large-scale 3D generative models. We introduce a token matching strategy that ensures accurate spatial correspondence during refinement, enabling local detail synthesis while preserving global structure. By carefully designing our training data to match the characteristics of synthesized coarse shapes, our method can effectively enhance shapes produced by various 3D generation and reconstruction approaches, from single-view to sparse multi-view inputs. Extensive experiments demonstrate that DetailGen3D achieves high-fidelity geometric detail synthesis while maintaining efficiency in training.

♻ ☆ IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations ICLR 2025

Zhibing Li, Tong Wu, Jing Tan, Mengchen Zhang, Jiaqi Wang, Dahua Lin

Capturing geometric and material information from images remains a fundamental challenge in computer vision and graphics. Traditional optimization-based methods often require hours of computational time to reconstruct geometry, material properties, and environmental lighting from dense multi-view inputs, while still struggling with inherent ambiguities between lighting and material. On the other hand, learning-based approaches leverage rich material priors from existing 3D object datasets but face challenges with maintaining multi-view consistency. In this paper, we introduce IDArb, a diffusion-based model designed to perform intrinsic decomposition on an arbitrary number of images under varying illuminations. Our method achieves accurate and multi-view consistent estimation on surface normals and material properties. This is made possible through a novel cross-view, cross-domain attention module and an illumination-augmented, view-adaptive training strategy. Additionally, we introduce ARB-Objaverse, a new dataset that provides large-scale multi-view intrinsic data and renderings under diverse lighting conditions, supporting robust training. Extensive experiments demonstrate that IDArb outperforms state-of-the-art methods both qualitatively and quantitatively. Moreover, our approach facilitates a range of downstream tasks, including single-image relighting, photometric stereo, and 3D reconstruction, highlighting its broad applications in realistic 3D content creation.

comment: ICLR 2025. Project Page: https://lizb6626.github.io/IDArb/

♻ ☆ Oriented Object Detection in Optical Remote Sensing Images using Deep Learning: A Survey

Kun Wang, Zi Wang, Zhang Li, Ang Su, Xichao Teng, Erting Pan, Minhao Liu, Qifeng Yu

Oriented object detection is one of the most fundamental and challenging tasks in remote sensing, aiming to locate and classify objects with arbitrary orientations. Recent advancements in deep learning have significantly enhanced the capabilities of oriented object detection. Given the rapid development of this field, this paper presents a comprehensive survey of recent advances in oriented object detection. To be specific, we begin by tracing the technical evolution from horizontal object detection to oriented object detection and highlighting the specific challenges, including feature misalignment, spatial misalignment, and oriented bounding box (OBB) regression problems. Subsequently, we further categorize existing methods into detection framework, OBB regression, and feature representations, and provide an in-depth discussion on how these approaches address the above challenges. In addition, we cover several publicly available datasets and evaluation protocols. Furthermore, we provide a comprehensive comparison and analysis of state-of-the-art methods. Toward the end of this paper, we identify several future directions for oriented object detection.

♻ ☆ Mixture of Experts Made Personalized: Federated Prompt Learning for Vision-Language Models ICLR 2025

Jun Luo, Chen Chen, Shandong Wu

Federated prompt learning benefits federated learning with CLIP-like Vision-Language Model's (VLM's) robust representation learning ability through prompt learning. However, current federated prompt learning methods are habitually restricted to the traditional FL paradigm, where the participating clients are generally only allowed to download a single globally aggregated model from the server. While justifiable for training full-sized models under federated settings, in this work, we argue that this paradigm is ill-suited for lightweight prompts. By facilitating the clients to download multiple pre-aggregated prompts as fixed non-local experts, we propose Personalized Federated Mixture of Adaptive Prompts (pFedMoAP), a novel FL framework that personalizes the prompt learning process through the lens of Mixture of Experts (MoE). pFedMoAP implements a local attention-based gating network that learns to generate enhanced text features for better alignment with local image data, benefiting from both local and downloaded non-local adaptive prompt experts. Extensive experiments on 9 datasets under various federated settings demonstrate the efficacy of the proposed pFedMoAP algorithm. The code is available at https://github.com/ljaiverson/pFedMoAP.

comment: ICLR 2025

♻ ☆ HCMA-UNet: A Hybrid CNN-Mamba UNet with Axial Self-Attention for Efficient Breast Cancer Segmentation

Haoxuan Li, Wei song, Peiwu Qin, Xi Yuan, Zhenglin Chen

Breast cancer lesion segmentation in DCE-MRI remains challenging due to heterogeneous tumor morphology and indistinct boundaries. To address these challenges, this study proposes a novel hybrid segmentation network, HCMA-UNet, for lesion segmentation of breast cancer. Our network consists of a lightweight CNN backbone and a Multi-view Axial Self-Attention Mamba (MISM) module. The MISM module integrates Visual State Space Block (VSSB) and Axial Self-Attention (ASA) mechanism, effectively reducing parameters through Asymmetric Split Channel (ASC) strategy to achieve efficient tri-directional feature extraction. Our lightweight model achieves superior performance with 2.87M parameters and 126.44 GFLOPs. A Feature-guided Region-aware loss function (FRLoss) is proposed to enhance segmentation accuracy. Extensive experiments on one private and two public DCE-MRI breast cancer datasets demonstrate that our approach achieves state-of-the-art performance while maintaining computational efficiency. FRLoss also exhibits good cross-architecture generalization capabilities. The source code is available at https://github.com/Haoxuanli-Thu/HCMA-UNet.

♻ ☆ HyperSIGMA: Hyperspectral Intelligence Comprehension Foundation Model

Di Wang, Meiqi Hu, Yao Jin, Yuchun Miao, Jiaqi Yang, Yichu Xu, Xiaolei Qin, Jiaqi Ma, Lingyu Sun, Chenxing Li, Chuan Fu, Hongruixuan Chen, Chengxi Han, Naoto Yokoya, Jing Zhang, Minqiang Xu, Lin Liu, Lefei Zhang, Chen Wu, Bo Du, Dacheng Tao, Liangpei Zhang

Accurate hyperspectral image (HSI) interpretation is critical for providing valuable insights into various earth observation-related applications such as urban planning, precision agriculture, and environmental monitoring. However, existing HSI processing methods are predominantly task-specific and scene-dependent, which severely limits their ability to transfer knowledge across tasks and scenes, thereby reducing the practicality in real-world applications. To address these challenges, we present HyperSIGMA, a vision transformer-based foundation model that unifies HSI interpretation across tasks and scenes, scalable to over one billion parameters. To overcome the spectral and spatial redundancy inherent in HSIs, we introduce a novel sparse sampling attention (SSA) mechanism, which effectively promotes the learning of diverse contextual features and serves as the basic block of HyperSIGMA. HyperSIGMA integrates spatial and spectral features using a specially designed spectral enhancement module. In addition, we construct a large-scale hyperspectral dataset, HyperGlobal-450K, for pre-training, which contains about 450K hyperspectral images, significantly surpassing existing datasets in scale. Extensive experiments on various high-level and low-level HSI tasks demonstrate HyperSIGMA's versatility and superior representational capability compared to current state-of-the-art methods. Moreover, HyperSIGMA shows significant advantages in scalability, robustness, cross-modal transferring capability, real-world applicability, and computational efficiency. The code and models will be released at https://github.com/WHU-Sigma/HyperSIGMA.

comment: Accepted by IEEE TPAMI. Project website: https://whu-sigma.github.io/HyperSIGMA

♻ ☆ Mind the GAP: Glimpse-based Active Perception improves generalization and sample efficiency of visual reasoning

Oleh Kolner, Thomas Ortner, Stanisław Woźniak, Angeliki Pantazi

Human capabilities in understanding visual relations are far superior to those of AI systems, especially for previously unseen objects. For example, while AI systems struggle to determine whether two such objects are visually the same or different, humans can do so with ease. Active vision theories postulate that the learning of visual relations is grounded in actions that we take to fixate objects and their parts by moving our eyes. In particular, the low-dimensional spatial information about the corresponding eye movements is hypothesized to facilitate the representation of relations between different image parts. Inspired by these theories, we develop a system equipped with a novel Glimpse-based Active Perception (GAP) that sequentially glimpses at the most salient regions of the input image and processes them at high resolution. Importantly, our system leverages the locations stemming from the glimpsing actions, along with the visual content around them, to represent relations between different parts of the image. The results suggest that the GAP is essential for extracting visual relations that go beyond the immediate visual content. Our approach reaches state-of-the-art performance on several visual reasoning tasks being more sample-efficient, and generalizing better to out-of-distribution visual inputs than prior models.

comment: 10 pages of main text and 8 pages appendices

♻ ☆ RedMotion: Motion Prediction via Redundancy Reduction

Royden Wagner, Omer Sahin Tas, Marvin Klemp, Carlos Fernandez, Christoph Stiller

We introduce RedMotion, a transformer model for motion prediction in self-driving vehicles that learns environment representations via redundancy reduction. Our first type of redundancy reduction is induced by an internal transformer decoder and reduces a variable-sized set of local road environment tokens, representing road graphs and agent data, to a fixed-sized global embedding. The second type of redundancy reduction is obtained by self-supervised learning and applies the redundancy reduction principle to embeddings generated from augmented views of road environments. Our experiments reveal that our representation learning approach outperforms PreTraM, Traj-MAE, and GraphDINO in a semi-supervised setting. Moreover, RedMotion achieves competitive results compared to HPTR or MTR++ in the Waymo Motion Prediction Challenge. Our open-source implementation is available at: https://github.com/kit-mrt/future-motion

comment: TMLR published version

♻ ☆ Fine-Grained Behavior and Lane Constraints Guided Trajectory Prediction Method

Wenyi Xiong, Jian Chen, Ziheng Qi

Trajectory prediction, as a critical component of autonomous driving systems, has attracted the attention of many researchers. Existing prediction algorithms focus on extracting more detailed scene features or selecting more reasonable trajectory destinations. However, in the face of dynamic and evolving future movements of the target vehicle, these algorithms cannot provide a fine-grained and continuous description of future behaviors and lane constraints, which degrades the prediction accuracy. To address this challenge, we present BLNet, a novel dualstream architecture that synergistically integrates behavioral intention recognition and lane constraint modeling through parallel attention mechanisms. The framework generates fine-grained behavior state queries (capturing spatial-temporal movement patterns) and lane queries (encoding lane topology constraints), supervised by two auxiliary losses, respectively. Subsequently, a two-stage decoder first produces trajectory proposals, then performs point-level refinement by jointly incorporating both the continuity of passed lanes and future motion features. Extensive experiments on two large datasets, nuScenes and Argoverse, show that our network exhibits significant performance gains over existing direct regression and goal-based algorithms.

comment: This work has been submitted to the IEEE for possible publication

♻ ☆ RePoseD: Efficient Relative Pose Estimation With Known Depth Information

Yaqing Ding, Viktor Kocur, Václav Vávra, Jian Yang, Torsten Sattler, Zuzana Kukelova

Recent advances in monocular depth estimation methods (MDE) and their improved accuracy open new possibilities for their applications. In this paper, we investigate how monocular depth estimates can be used for relative pose estimation. In particular, we are interested in answering the question whether using MDEs improves results over traditional point-based methods. We propose a novel framework for estimating the relative pose of two cameras from point correspondences with associated monocular depths. Since depth predictions are typically defined up to an unknown scale or even both unknown scale and shift parameters, our solvers jointly estimate the scale or both the scale and shift parameters along with the relative pose. We derive efficient solvers considering different types of depths for three camera configurations: (1) two calibrated cameras, (2) two cameras with an unknown shared focal length, and (3) two cameras with unknown different focal lengths. Our new solvers outperform state-of-the-art depth-aware solvers in terms of speed and accuracy. In extensive real experiments on multiple datasets and with various MDEs, we discuss which depth-aware solvers are preferable in which situation. The code will be made publicly available.

comment: 18 pages

♻ ☆ Enhanced Controllability of Diffusion Models via Feature Disentanglement and Realism-Enhanced Sampling Methods ECCV 2024

Wonwoong Cho, Hareesh Ravi, Midhun Harikumar, Vinh Khuc, Krishna Kumar Singh, Jingwan Lu, David I. Inouye, Ajinkya Kale

As Diffusion Models have shown promising performance, a lot of efforts have been made to improve the controllability of Diffusion Models. However, how to train Diffusion Models to have the disentangled latent spaces and how to naturally incorporate the disentangled conditions during the sampling process have been underexplored. In this paper, we present a training framework for feature disentanglement of Diffusion Models (FDiff). We further propose two sampling methods that can boost the realism of our Diffusion Models and also enhance the controllability. Concisely, we train Diffusion Models conditioned on two latent features, a spatial content mask, and a flattened style embedding. We rely on the inductive bias of the denoising process of Diffusion Models to encode pose/layout information in the content feature and semantic/style information in the style feature. Regarding the sampling methods, we first generalize Composable Diffusion Models (GCDM) by breaking the conditional independence assumption to allow for some dependence between conditional inputs, which is shown to be effective in realistic generation in our experiments. Second, we propose timestep-dependent weight scheduling for content and style features to further improve the performance. We also observe better controllability of our proposed methods compared to existing methods in image manipulation and image translation.

comment: ECCV 2024; Code will be opened after a patent application is granted

♻ ☆ Att-Adapter: A Robust and Precise Domain-Specific Multi-Attributes T2I Diffusion Adapter via Conditional Variational Autoencoder

Wonwoong Cho, Yan-Ying Chen, Matthew Klenk, David I. Inouye, Yanxia Zhang

Text-to-Image (T2I) Diffusion Models have achieved remarkable performance in generating high quality images. However, enabling precise control of continuous attributes, especially multiple attributes simultaneously, in a new domain (e.g., numeric values like eye openness or car width) with text-only guidance remains a significant challenge. To address this, we introduce the Attribute (Att) Adapter, a novel plug-and-play module designed to enable fine-grained, multi-attributes control in pretrained diffusion models. Our approach learns a single control adapter from a set of sample images that can be unpaired and contain multiple visual attributes. The Att-Adapter leverages the decoupled cross attention module to naturally harmonize the multiple domain attributes with text conditioning. We further introduce Conditional Variational Autoencoder (CVAE) to the Att-Adapter to mitigate overfitting, matching the diverse nature of the visual world. Evaluations on two public datasets show that Att-Adapter outperforms all LoRA-based baselines in controlling continuous attributes. Additionally, our method enables a broader control range and also improves disentanglement across multiple attributes, surpassing StyleGAN-based techniques. Notably, Att-Adapter is flexible, requiring no paired synthetic data for training, and is easily scalable to multiple attributes within a single model.

♻ ☆ MSCMNet: Multi-scale Semantic Correlation Mining for Visible-Infrared Person Re-Identification

Xuecheng Hua, Ke Cheng, Hu Lu, Juanjuan Tu, Yuanquan Wang, Shitong Wang

The main challenge in the Visible-Infrared Person Re-Identification (VI-ReID) task lies in how to extract discriminative features from different modalities for matching purposes. While the existing well works primarily focus on minimizing the modal discrepancies, the modality information can not thoroughly be leveraged. To solve this problem, a Multi-scale Semantic Correlation Mining network (MSCMNet) is proposed to comprehensively exploit semantic features at multiple scales and simultaneously reduce modality information loss as small as possible in feature extraction. The proposed network contains three novel components. Firstly, after taking into account the effective utilization of modality information, the Multi-scale Information Correlation Mining Block (MIMB) is designed to explore semantic correlations across multiple scales. Secondly, in order to enrich the semantic information that MIMB can utilize, a quadruple-stream feature extractor (QFE) with non-shared parameters is specifically designed to extract information from different dimensions of the dataset. Finally, the Quadruple Center Triplet Loss (QCT) is further proposed to address the information discrepancy in the comprehensive features. Extensive experiments on the SYSU-MM01, RegDB, and LLCM datasets demonstrate that the proposed MSCMNet achieves the greatest accuracy.

♻ ☆ Efficient Semantic Segmentation via Lightweight Multiple-Information Interaction Network

Yangyang Qiu, Guoan Xu, Guangwei Gao, Zhenhua Guo, Yi Yu, Chia-Wen Lin

Recently, integrating the local modeling capabilities of Convolutional Neural Networks (CNNs) with the global dependency strengths of Transformers has created a sensation in the semantic segmentation community. However, substantial computational workloads and high hardware memory demands remain major obstacles to their further application in real-time scenarios. In this work, we propose a Lightweight Multiple-Information Interaction Network (LMIINet) for real-time semantic segmentation, which effectively combines CNNs and Transformers while reducing redundant computations and memory footprints. It features Lightweight Feature Interaction Bottleneck (LFIB) modules comprising efficient convolutions that enhance context integration. Additionally, improvements are made to the Flatten Transformer by enhancing local and global feature interaction to capture detailed semantic information. Incorporating a combination coefficient learning scheme in both LFIB and Transformer blocks facilitates improved feature interaction. Extensive experiments demonstrate that LMIINet excels in balancing accuracy and efficiency. With only 0.72M parameters and 11.74G FLOPs (Floating Point Operations Per Second), LMIINet achieves 72.0\% mIoU at 100 FPS (Frames Per Second) on the Cityscapes test set and 69.94\% mIoU (mean Intersection over Union) at 160 FPS on the CamVid test dataset using a single RTX2080Ti GPU.

comment: 10 pages, 6 figures, 9 tables

♻ ☆ A Comparative Study of Scanpath Models in Graph-Based Visualization

Angela Lopez-Cardona, Parvin Emami, Sebastian Idesis, Saravanakumar Duraisamy, Luis A. Leiva, Ioannis Arapakis

Information Visualization (InfoVis) systems utilize visual representations to enhance data interpretation. Understanding how visual attention is allocated is essential for optimizing interface design. However, collecting Eye-tracking (ET) data presents challenges related to cost, privacy, and scalability. Computational models provide alternatives for predicting gaze patterns, thereby advancing InfoVis research. In our study, we conducted an ET experiment with 40 participants who analyzed graphs while responding to questions of varying complexity within the context of digital forensics. We compared human scanpaths with synthetic ones generated by models such as DeepGaze, UMSS, and Gazeformer. Our research evaluates the accuracy of these models and examines how question complexity and number of nodes influence performance. This work contributes to the development of predictive modeling in visual analytics, offering insights that can enhance the design and effectiveness of InfoVis systems.

♻ ☆ ConsistencyDet: A Few-step Denoising Framework for Object Detection Using the Consistency Model

Lifan Jiang, Zhihui Wang, Changmiao Wang, Ming Li, Jiaxu Leng, Xindong Wu

Object detection, a quintessential task in the realm of perceptual computing, can be tackled using a generative methodology. In the present study, we introduce a novel framework designed to articulate object detection as a denoising diffusion process, which operates on the perturbed bounding boxes of annotated entities. This framework, termed \textbf{ConsistencyDet}, leverages an innovative denoising concept known as the Consistency Model. The hallmark of this model is its self-consistency feature, which empowers the model to map distorted information from any time step back to its pristine state, thereby realizing a \textbf{``few-step denoising''} mechanism. Such an attribute markedly elevates the operational efficiency of the model, setting it apart from the conventional Diffusion Model. Throughout the training phase, ConsistencyDet initiates the diffusion sequence with noise-infused boxes derived from the ground-truth annotations and conditions the model to perform the denoising task. Subsequently, in the inference stage, the model employs a denoising sampling strategy that commences with bounding boxes randomly sampled from a normal distribution. Through iterative refinement, the model transforms an assortment of arbitrarily generated boxes into definitive detections. Comprehensive evaluations employing standard benchmarks, such as MS-COCO and LVIS, corroborate that ConsistencyDet surpasses other leading-edge detectors in performance metrics. Our code is available at https://anonymous.4open.science/r/ConsistencyDet-37D5.

♻ ☆ SVInvNet: A Densely Connected Encoder-Decoder Architecture for Seismic Velocity Inversion

Mojtaba Najafi Khatounabad, Hacer Yalim Keles, Selma Kadioglu

This study presents a deep learning-based approach to seismic velocity inversion problem, focusing on both noisy and noiseless training datasets of varying sizes. Our Seismic Velocity Inversion Network (SVInvNet) introduces a novel architecture that contains a multi-connection encoder-decoder structure enhanced with dense blocks. This design is specifically tuned to effectively process time series data, which is essential for addressing the challenges of non-linear seismic velocity inversion. For training and testing, we created diverse seismic velocity models, including multi-layered, faulty, and salt dome categories. We also investigated how different kinds of ambient noise, both coherent and stochastic, and the size of the training dataset affect learning outcomes. SVInvNet is trained on datasets ranging from 750 to 6,000 samples and is tested using a large benchmark dataset of 12,000 samples. Despite its fewer parameters compared to the baseline model, SVInvNet achieves superior performance with this dataset. The performance of SVInvNet was further evaluated using the OpenFWI dataset and Marmousi-derived velocity models. The comparative analysis clearly reveals the effectiveness of the proposed model.

comment: This is the preprint of the accepted manuscript to appear in IEEE Transactions on Geoscience and Remote Sensing

♻ ☆ Self-Supervised Pretraining for Aerial Road Extraction

Rupert Polley, Sai Vignesh Abishek Deenadayalan, J. Marius Zöllner

Deep neural networks for aerial image segmentation require large amounts of labeled data, but high-quality aerial datasets with precise annotations are scarce and costly to produce. To address this limitation, we propose a self-supervised pretraining method that improves segmentation performance while reducing reliance on labeled data. Our approach uses inpainting-based pretraining, where the model learns to reconstruct missing regions in aerial images, capturing their inherent structure before being fine-tuned for road extraction. This method improves generalization, enhances robustness to domain shifts, and is invariant to model architecture and dataset choice. Experiments show that our pretraining significantly boosts segmentation accuracy, especially in low-data regimes, making it a scalable solution for aerial image analysis.

comment: Accepted at 36th IEEE Intelligent Vehicles Symposium (IV) 2025 Joint Workshop on Safety, Metrics and Benchmarks for Autonomous Driving

♻ ☆ DG-TTA: Out-of-domain Medical Image Segmentation through Augmentation and Descriptor-driven Domain Generalization and Test-Time Adaptation

Christian Weihsbach, Christian N. Kruse, Alexander Bigalke, Mattias P. Heinrich

Purpose: Applying pre-trained medical deep learning segmentation models on out-of-domain images often yields predictions of insufficient quality. In this study, we propose to use a powerful generalizing descriptor along with augmentation to enable domain-generalized pre-training and test-time adaptation, achieving high-quality segmentation in unseen domains. Materials and Methods: In this retrospective study five different publicly available datasets (2012 to 2022) including 3D CT and MRI images are used to evaluate segmentation performance in out-of-domain scenarios. The settings include abdominal, spine, and cardiac imaging. The data is randomly split into training and test samples. Domain-generalized pre-training on source data is used to obtain the best initial performance in the target domain. We introduce the combination of the generalizing SSC descriptor and GIN intensity augmentation for optimal generalization. Segmentation results are subsequently optimized at test time, where we propose to adapt the pre-trained models for every unseen scan with a consistency scheme using the same augmentation-descriptor combination. The segmentation is evaluated using Dice similarity and Hausdorff distance and the significance of improvements is tested with the Wilcoxon signed-rank test. Results: The proposed generalized pre-training and subsequent test-time adaptation improves model performance significantly in CT to MRI cross-domain prediction for abdominal (+46.2% and +28.2% Dice), spine (+72.9%), and cardiac (+14.2% and +55.7% Dice) scenarios (p<0.001). Conclusion: Our method enables optimal, independent usage of medical image source and target data and bridges domain gaps successfully with a compact and efficient methodology. Open-source code available at: https://github.com/multimodallearning/DG-TTA

♻ ☆ Nonhuman Primate Brain Tissue Segmentation Using a Transfer Learning Approach

Zhen Lin, Hongyu Yuan, Richard Barcus, Qing Lyu, Sucheta Chakravarty, Megan E. Lipford, Carol A. Shively, Suzanne Craft, Mohammad Kawas, Jeongchul Kim, Christopher T. Whitlow

Non-human primates (NHPs) serve as critical models for understanding human brain function and neurological disorders due to their close evolutionary relationship with humans. Accurate brain tissue segmentation in NHPs is critical for understanding neurological disorders, but challenging due to the scarcity of annotated NHP brain MRI datasets, the small size of the NHP brain, the limited resolution of available imaging data and the anatomical differences between human and NHP brains. To address these challenges, we propose a novel approach utilizing STU-Net with transfer learning to leverage knowledge transferred from human brain MRI data to enhance segmentation accuracy in the NHP brain MRI, particularly when training data is limited. The combination of STU-Net and transfer learning effectively delineates complex tissue boundaries and captures fine anatomical details specific to NHP brains. Notably, our method demonstrated improvement in segmenting small subcortical structures such as putamen and thalamus that are challenging to resolve with limited spatial resolution and tissue contrast, and achieved DSC of over 0.88, IoU over 0.8 and HD95 under 7. This study introduces a robust method for multi-class brain tissue segmentation in NHPs, potentially accelerating research in evolutionary neuroscience and preclinical studies of neurological disorders relevant to human health.

♻ ☆ Exploring Scene Affinity for Semi-Supervised LiDAR Semantic Segmentation CVPR2025

Chuandong Liu, Xingxing Weng, Shuguo Jiang, Pengcheng Li, Lei Yu, Gui-Song Xia

This paper explores scene affinity (AIScene), namely intra-scene consistency and inter-scene correlation, for semi-supervised LiDAR semantic segmentation in driving scenes. Adopting teacher-student training, AIScene employs a teacher network to generate pseudo-labeled scenes from unlabeled data, which then supervise the student network's learning. Unlike most methods that include all points in pseudo-labeled scenes for forward propagation but only pseudo-labeled points for backpropagation, AIScene removes points without pseudo-labels, ensuring consistency in both forward and backward propagation within the scene. This simple point erasure strategy effectively prevents unsupervised, semantically ambiguous points (excluded in backpropagation) from affecting the learning of pseudo-labeled points. Moreover, AIScene incorporates patch-based data augmentation, mixing multiple scenes at both scene and instance levels. Compared to existing augmentation techniques that typically perform scene-level mixing between two scenes, our method enhances the semantic diversity of labeled (or pseudo-labeled) scenes, thereby improving the semi-supervised performance of segmentation models. Experiments show that AIScene outperforms previous methods on two popular benchmarks across four settings, achieving notable improvements of 1.9% and 2.1% in the most challenging 1% labeled data.

comment: Accepted by CVPR2025

♻ ☆ Introducing the Short-Time Fourier Kolmogorov Arnold Network: A Dynamic Graph CNN Approach for Tree Species Classification in 3D Point Clouds

Said Ohamouddou, Mohamed Ohamouddou, Hanaa El Afia, Abdellatif El Afia, Rafik Lasri, Raddouane Chiheb

Accurate classification of tree species based on Terrestrial Laser Scanning (TLS) and Airborne Laser Scanning (ALS) is essential for biodiversity conservation. While advanced deep learning models for 3D point cloud classification have demonstrated strong performance in this domain, their high complexity often hinders the development of efficient, low-computation architectures. In this paper, we introduce STFT-KAN, a novel Kolmogorov-Arnold network that integrates the Short-Time Fourier Transform (STFT), which can replace the standard linear layer with activation. We implemented STFT-KAN within a lightweight version of DGCNN, called liteDGCNN, to classify tree species using the TLS data. Our experiments show that STFT-KAN outperforms existing KAN variants by effectively balancing model complexity and performance with parameter count reduction, achieving competitive results compared to MLP-based models. Additionally, we evaluated a hybrid architecture that combines MLP in edge convolution with STFT-KAN in other layers, achieving comparable performance to MLP models while reducing the parameter count by 50% and 75% compared to other KAN-based variants. Furthermore, we compared our model to leading 3D point cloud learning approaches, demonstrating that STFT-KAN delivers competitive results compared to the state-of-the-art method PointMLP lite with an 87% reduction in parameter count.

♻ ☆ Exploiting Mixture-of-Experts Redundancy Unlocks Multimodal Generative Abilities

Raman Dutt, Harleen Hanspal, Guoxuan Xia, Petru-Daniel Tudosiu, Alexander Black, Yongxin Yang, Steven McDonagh, Sarah Parisot

In this work, we undertake the challenge of augmenting the existing generative capabilities of pre-trained text-only large language models (LLMs) with multi-modal generation capability while satisfying two core constraints: C1 preserving the preservation of original language generative capabilities with negligible performance degradation, and C2 adhering to a small parameter budget to learn the new modality, ensuring scalability and efficiency. In contrast to current approaches that add dedicated modules, thereby significantly increasing the parameter count, we propose a method that leverages the underutilized capacity inherent in deep models. Specifically, we exploit the parameter redundancy within Mixture-of-Experts (MoEs) as a source of additional capacity for learning a new modality, enabling better parameter efficiency (C1). Moreover, we preserve the original language generation capabilities by applying low-rank adaptation exclusively to the tokens of the new modality (C2). Furthermore, we introduce a novel parameter initialization scheme based on the Gromov-Wasserstein distance to improve convergence and training stability. Through an extensive analysis of the routing mechanism, we uncover the emergence of modality-specific pathways and decreased redundancy within the experts that can efficiently unlock multi-modal generative capabilities. Overall, our method can be seamlessly applied to a wide range of contemporary LLMs, providing a new pathway for transitioning from uni-modal to multi-modal architectures.

♻ ☆ DoubleDiffusion: Combining Heat Diffusion with Denoising Diffusion for Texture Generation on 3D Meshes

Xuyang Wang, Ziang Cheng, Zhenyu Li, Jiayu Yang, Haorui Ji, Pan Ji, Mehrtash Harandi, Richard Hartley, Hongdong Li

This paper addresses the problem of generating textures for 3D mesh assets. Existing approaches often rely on image diffusion models to generate multi-view image observations, which are then transformed onto the mesh surface to produce a single texture. However, due to the gap between multi-view images and 3D space, such process is susceptible to arange of issues such as geometric inconsistencies, visibility occlusion, and baking artifacts. To overcome this problem, we propose a novel approach that directly generates texture on 3D meshes. Our approach leverages heat dissipation diffusion, which serves as an efficient operator that propagates features on the geometric surface of a mesh, while remaining insensitive to the specific layout of the wireframe. By integrating this technique into a generative diffusion pipeline, we significantly improve the efficiency of texture generation compared to existing texture generation methods. We term our approach DoubleDiffusion, as it combines heat dissipation diffusion with denoising diffusion to enable native generative learning on 3D mesh surfaces.

comment: Codes: https://github.com/Wxyxixixi/DoubleDiffusion_3D_Mesh

♻ ☆ Attention-Guided Multi-scale Interaction Network for Face Super-Resolution

Xujie Wan, Wenjie Li, Guangwei Gao, Huimin Lu, Jian Yang, Chia-Wen Lin

Recently, CNN and Transformer hybrid networks demonstrated excellent performance in face super-resolution (FSR) tasks. Since numerous features at different scales in hybrid networks, how to fuse these multi-scale features and promote their complementarity is crucial for enhancing FSR. However, existing hybrid network-based FSR methods ignore this, only simply combining the Transformer and CNN. To address this issue, we propose an attention-guided Multi-scale interaction network (AMINet), which contains local and global feature interactions and encoder-decoder phase feature interactions. Specifically, we propose a Local and Global Feature Interaction Module (LGFI) to promote fusions of global features and different receptive fields' local features extracted by our Residual Depth Feature Extraction Module (RDFE). Additionally, we propose a Selective Kernel Attention Fusion Module (SKAF) to adaptively select fusions of different features within LGFI and encoder-decoder phases. Our above design allows the free flow of multi-scale features from within modules and between encoder and decoder, which can promote the complementarity of different scale features to enhance FSR. Comprehensive experiments confirm that our method consistently performs well with less computational consumption and faster inference.

comment: 13 pages, 11 figures, 10 tables

♻ ☆ UniGS: Modeling Unitary 3D Gaussians for Novel View Synthesis from Sparse-view Images

Jiamin Wu, Kenkun Liu, Yukai Shi, Xiaoke Jiang, Yuan Yao, Lei Zhang

In this work, we introduce UniGS, a novel 3D Gaussian reconstruction and novel view synthesis model that predicts a high-fidelity representation of 3D Gaussians from arbitrary number of posed sparse-view images. Previous methods often regress 3D Gaussians locally on a per-pixel basis for each view and then transfer them to world space and merge them through point concatenation. In contrast, Our approach involves modeling unitary 3D Gaussians in world space and updating them layer by layer. To leverage information from multi-view inputs for updating the unitary 3D Gaussians, we develop a DETR (DEtection TRansformer)-like framework, which treats 3D Gaussians as queries and updates their parameters by performing multi-view cross-attention (MVDFA) across multiple input images, which are treated as keys and values. This approach effectively avoids `ghosting' issue and allocates more 3D Gaussians to complex regions. Moreover, since the number of 3D Gaussians used as decoder queries is independent of the number of input views, our method allows arbitrary number of multi-view images as input without causing memory explosion or requiring retraining. Extensive experiments validate the advantages of our approach, showcasing superior performance over existing methods quantitatively (improving PSNR by 4.2 dB when trained on Objaverse and tested on the GSO benchmark) and qualitatively. The code will be released at https://github.com/jwubz123/UNIG.

♻ ☆ Image as an IMU: Estimating Camera Motion from a Single Motion-Blurred Image

Jerred Chen, Ronald Clark

In many robotics and VR/AR applications, fast camera motions cause a high level of motion blur, causing existing camera pose estimation methods to fail. In this work, we propose a novel framework that leverages motion blur as a rich cue for motion estimation rather than treating it as an unwanted artifact. Our approach works by predicting a dense motion flow field and a monocular depth map directly from a single motion-blurred image. We then recover the instantaneous camera velocity by solving a linear least squares problem under the small motion assumption. In essence, our method produces an IMU-like measurement that robustly captures fast and aggressive camera movements. To train our model, we construct a large-scale dataset with realistic synthetic motion blur derived from ScanNet++v2 and further refine our model by training end-to-end on real data using our fully differentiable pipeline. Extensive evaluations on real-world benchmarks demonstrate that our method achieves state-of-the-art angular and translational velocity estimates, outperforming current methods like MASt3R and COLMAP.

comment: Project page: https://jerredchen.github.io/image-as-imu/

♻ ☆ Think or Not Think: A Study of Explicit Thinking inRule-Based Visual Reinforcement Fine-Tuning

Ming Li, Jike Zhong, Shitian Zhao, Yuxiang Lai, Kaipeng Zhang

This paper investigates rule-based reinforcement learning (RL) fine-tuning for visual classification using multi-modal large language models (MLLMs) and the role of the thinking process. We begin by exploring \textit{CLS-RL}, a method that leverages verifiable signals as rewards to encourage MLLMs to 'think' before classifying. Our experiments across \textbf{eleven} datasets demonstrate that CLS-RL achieves significant improvements over supervised fine-tuning (SFT) in both base-to-new generalization and few-shot learning scenarios. Notably, we observe a 'free-lunch' phenomenon where fine-tuning on one dataset unexpectedly enhances performance on others, suggesting that RL effectively teaches fundamental classification skills. However, we question whether the explicit thinking, a critical aspect of rule-based RL, is always beneficial or indispensable. Challenging the conventional assumption that complex reasoning enhances performance, we introduce \textit{No-Thinking-RL}, a novel approach that minimizes the model's thinking during fine-tuning by utilizing an equality accuracy reward. Our experiments reveal that No-Thinking-RL achieves superior in-domain performance and generalization capabilities compared to CLS-RL, while requiring significantly less fine-tuning time. This underscores that, contrary to prevailing assumptions, reducing the thinking process can lead to more efficient and effective MLLM fine-tuning for some visual tasks. Furthermore, No-Thinking-RL demonstrates enhanced performance on other visual benchmarks, such as a 6.4\% improvement on CVBench. We hope our findings provides insights into the impact of thinking in RL-based fine-tuning.

comment: Preprint, work in progress. Add results on CVBench

♻ ☆ PhyT2V: LLM-Guided Iterative Self-Refinement for Physics-Grounded Text-to-Video Generation

Qiyao Xue, Xiangyu Yin, Boyuan Yang, Wei Gao

Text-to-video (T2V) generation has been recently enabled by transformer-based diffusion models, but current T2V models lack capabilities in adhering to the real-world common knowledge and physical rules, due to their limited understanding of physical realism and deficiency in temporal modeling. Existing solutions are either data-driven or require extra model inputs, but cannot be generalizable to out-of-distribution domains. In this paper, we present PhyT2V, a new data-independent T2V technique that expands the current T2V model's capability of video generation to out-of-distribution domains, by enabling chain-of-thought and step-back reasoning in T2V prompting. Our experiments show that PhyT2V improves existing T2V models' adherence to real-world physical rules by 2.3x, and achieves 35% improvement compared to T2V prompt enhancers. The source codes are available at: https://github.com/pittisl/PhyT2V.

comment: 28 pages

♻ ☆ FisherTune: Fisher-Guided Robust Tuning of Vision Foundation Models for Domain Generalized Segmentation

Dong Zhao, Jinlong Li, Shuang Wang, Mengyao Wu, Qi Zang, Nicu Sebe, Zhun Zhong

Vision Foundation Models (VFMs) excel in generalization due to large-scale pretraining, but fine-tuning them for Domain Generalized Semantic Segmentation (DGSS) while maintaining this ability remains challenging. Existing approaches either selectively fine-tune parameters or freeze the VFMs and update only the adapters, both of which may underutilize the VFMs' full potential in DGSS tasks. We observe that domain-sensitive parameters in VFMs, arising from task and distribution differences, can hinder generalization. To address this, we propose \textbf{FisherTune}, a robust fine-tuning method guided by the Domain-Related Fisher Information Matrix (DR-FIM). DR-FIM measures parameter sensitivity across tasks and domains, enabling selective updates that preserve generalization and enhance DGSS adaptability. FisherTune incorporates variational inference to stabilize DR-FIM estimation, treating parameters as Gaussian-distributed variables and leveraging pre-trained priors. Extensive experiments show that FisherTune achieves superior cross-domain segmentation while maintaining generalization, outperforming selective-parameter and adapter-based methods.

♻ ☆ Lie Detector: Unified Backdoor Detection via Cross-Examination Framework

Xuan Wang, Siyuan Liang, Dongping Liao, Han Fang, Aishan Liu, Xiaochun Cao, Yu-liang Lu, Ee-Chien Chang, Xitong Gao

Institutions with limited data and computing resources often outsource model training to third-party providers in a semi-honest setting, assuming adherence to prescribed training protocols with pre-defined learning paradigm (e.g., supervised or semi-supervised learning). However, this practice can introduce severe security risks, as adversaries may poison the training data to embed backdoors into the resulting model. Existing detection approaches predominantly rely on statistical analyses, which often fail to maintain universally accurate detection accuracy across different learning paradigms. To address this challenge, we propose a unified backdoor detection framework in the semi-honest setting that exploits cross-examination of model inconsistencies between two independent service providers. Specifically, we integrate central kernel alignment to enable robust feature similarity measurements across different model architectures and learning paradigms, thereby facilitating precise recovery and identification of backdoor triggers. We further introduce backdoor fine-tuning sensitivity analysis to distinguish backdoor triggers from adversarial perturbations, substantially reducing false positives. Extensive experiments demonstrate that our method achieves superior detection performance, improving accuracy by 5.4%, 1.6%, and 11.9% over SoTA baselines across supervised, semi-supervised, and autoregressive learning tasks, respectively. Notably, it is the first to effectively detect backdoors in multimodal large language models, further highlighting its broad applicability and advancing secure deep learning.

♻ ☆ An End-to-End Robust Point Cloud Semantic Segmentation Network with Single-Step Conditional Diffusion Models

Wentao Qu, Jing Wang, YongShun Gong, Xiaoshui Huang, Liang Xiao

Existing conditional Denoising Diffusion Probabilistic Models (DDPMs) with a Noise-Conditional Framework (NCF) remain challenging for 3D scene understanding tasks, as the complex geometric details in scenes increase the difficulty of fitting the gradients of the data distribution (the scores) from semantic labels. This also results in longer training and inference time for DDPMs compared to non-DDPMs. From a different perspective, we delve deeply into the model paradigm dominated by the Conditional Network. In this paper, we propose an end-to-end robust semantic Segmentation Network based on a Conditional-Noise Framework (CNF) of DDPMs, named CDSegNet. Specifically, CDSegNet models the Noise Network (NN) as a learnable noise-feature generator. This enables the Conditional Network (CN) to understand 3D scene semantics under multi-level feature perturbations, enhancing the generalization in unseen scenes. Meanwhile, benefiting from the noise system of DDPMs, CDSegNet exhibits strong noise and sparsity robustness in experiments. Moreover, thanks to CNF, CDSegNet can generate the semantic labels in a single-step inference like non-DDPMs, due to avoiding directly fitting the scores from semantic labels in the dominant network of CDSegNet. On public indoor and outdoor benchmarks, CDSegNet significantly outperforms existing methods, achieving state-of-the-art performance.

♻ ☆ OncoReg: Medical Image Registration for Oncological Challenges

Wiebke Heyer, Yannic Elser, Lennart Berkel, Xinrui Song, Xuanang Xu, Pingkun Yan, Xi Jia, Jinming Duan, Zi Li, Tony C. W. Mok, BoWen LI, Christian Staackmann, Christoph Großbröhmer, Lasse Hansen, Alessa Hering, Malte M. Sieren, Mattias P. Heinrich

In modern cancer research, the vast volume of medical data generated is often underutilised due to challenges related to patient privacy. The OncoReg Challenge addresses this issue by enabling researchers to develop and validate image registration methods through a two-phase framework that ensures patient privacy while fostering the development of more generalisable AI models. Phase one involves working with a publicly available dataset, while phase two focuses on training models on a private dataset within secure hospital networks. OncoReg builds upon the foundation established by the Learn2Reg Challenge by incorporating the registration of interventional cone-beam computed tomography (CBCT) with standard planning fan-beam CT (FBCT) images in radiotherapy. Accurate image registration is crucial in oncology, particularly for dynamic treatment adjustments in image-guided radiotherapy, where precise alignment is necessary to minimise radiation exposure to healthy tissues while effectively targeting tumours. This work details the methodology and data behind the OncoReg Challenge and provides a comprehensive analysis of the competition entries and results. Findings reveal that feature extraction plays a pivotal role in this registration task. A new method emerging from this challenge demonstrated its versatility, while established approaches continue to perform comparably to newer techniques. Both deep learning and classical approaches still play significant roles in image registration, with the combination of methods - particularly in feature extraction - proving most effective.

comment: 26 pages, 6 figures

♻ ☆ MambaPEFT: Exploring Parameter-Efficient Fine-Tuning for Mamba ICLR2025

Masakazu Yoshimura, Teruaki Hayashi, Yota Maeda

An ecosystem of Transformer-based models has been established by building large models with extensive data. Parameter-efficient fine-tuning (PEFT) is a crucial technology for deploying these models to downstream tasks with minimal cost while achieving effective performance. Recently, Mamba, a State Space Model (SSM)-based model, has attracted attention as a potential alternative to Transformers. While many large-scale Mamba-based models have been proposed, efficiently adapting pre-trained Mamba-based models to downstream tasks remains unexplored. In this paper, we conduct an exploratory analysis of PEFT methods for Mamba. We investigate the effectiveness of existing PEFT methods for Transformers when applied to Mamba. We also modify these methods to better align with the Mamba architecture. Additionally, we propose new Mamba-specific PEFT methods that leverage the distinctive structure of Mamba. Our experiments indicate that PEFT performs more effectively for Mamba than Transformers. Lastly, we demonstrate how to effectively combine multiple PEFT methods and provide a framework that outperforms previous works. To ensure reproducibility, we will release the code after publication.

comment: Accepted to ICLR2025

♻ ☆ Stable-Makeup: When Real-World Makeup Transfer Meets Diffusion Model

Yuxuan Zhang, Yirui Yuan, Yiren Song, Jiaming Liu

Current makeup transfer methods are limited to simple makeup styles, making them difficult to apply in real-world scenarios. In this paper, we introduce Stable-Makeup, a novel diffusion-based makeup transfer method capable of robustly transferring a wide range of real-world makeup, onto user-provided faces. Stable-Makeup is based on a pre-trained diffusion model and utilizes a Detail-Preserving (D-P) makeup encoder to encode makeup details. It also employs content and structural control modules to preserve the content and structural information of the source image. With the aid of our newly added makeup cross-attention layers in U-Net, we can accurately transfer the detailed makeup to the corresponding position in the source image. After content-structure decoupling training, Stable-Makeup can maintain content and the facial structure of the source image. Moreover, our method has demonstrated strong robustness and generalizability, making it applicable to varioustasks such as cross-domain makeup transfer, makeup-guided text-to-image generation and so on. Extensive experiments have demonstrated that our approach delivers state-of-the-art (SOTA) results among existing makeup transfer methods and exhibits a highly promising with broad potential applications in various related fields. Code released: https://github.com/Xiaojiu-z/Stable-Makeup

♻ ☆ Local Information Matters: Inference Acceleration For Grounded Conversation Generation Models Through Adaptive Local-Aware Token Pruning

Bizhe Bai, Jianjian Cao, Yadan Luo, Tao Chen

Grounded Conversation Generation (GCG) is an emerging vision-language task that requires models to generate natural language responses seamlessly intertwined with corresponding object segmentation masks. Recent models, such as GLaMM and OMG-LLaVA, achieve pixel-level grounding but incur significant computational costs due to processing a large number of visual tokens. Existing token pruning methods, like FastV and PyramidDrop, fail to preserve the local visual features critical for accurate grounding, leading to substantial performance drops in GCG tasks. To address this, we propose Adaptive Local-Aware Token Pruning (ALTP), a simple yet effective framework that accelerates GCG models by prioritizing local object information. ALTP introduces two key components: (1) Detail Density Capture (DDC), which uses superpixel segmentation to retain tokens in object-centric regions, preserving fine-grained details, and (2) Dynamic Density Formation (DDF), which dynamically allocates tokens based on information density, ensuring higher retention in semantically rich areas. Extensive experiments on the GranDf dataset demonstrate that ALTP significantly outperforms existing token pruning methods, such as FastV and PyramidDrop, on both GLaMM and OMG-LLaVA models. Notably, when applied to GLaMM, ALTP achieves a 90% reduction in visual tokens with a 4.9% improvement in AP50 and a 5.0% improvement in Recall compared to PyramidDrop. Similarly, on OMG-LLaVA, ALTP improves AP by 2.1% and mIOU by 3.0% at a 90% token reduction compared with PDrop.

♻ ☆ Mr. DETR: Instructive Multi-Route Training for Detection Transformers CVPR 2025

Chang-Bin Zhang, Yujie Zhong, Kai Han

Existing methods enhance the training of detection transformers by incorporating an auxiliary one-to-many assignment. In this work, we treat the model as a multi-task framework, simultaneously performing one-to-one and one-to-many predictions. We investigate the roles of each component in the transformer decoder across these two training targets, including self-attention, cross-attention, and feed-forward network. Our empirical results demonstrate that any independent component in the decoder can effectively learn both targets simultaneously, even when other components are shared. This finding leads us to propose a multi-route training mechanism, featuring a primary route for one-to-one prediction and two auxiliary training routes for one-to-many prediction. We enhance the training mechanism with a novel instructive self-attention that dynamically and flexibly guides object queries for one-to-many prediction. The auxiliary routes are removed during inference, ensuring no impact on model architecture or inference cost. We conduct extensive experiments on various baselines, achieving consistent improvements as shown in Figure 1. Project page: https://visual-ai.github.io/mrdetr

comment: Accepted by CVPR 2025, Project page: https://visual-ai.github.io/mrdetr

♻ ☆ ControlSR: Taming Diffusion Models for Consistent Real-World Image Super Resolution

Yuhao Wan, Peng-Tao Jiang, Qibin Hou, Hao Zhang, Jinwei Chen, Ming-Ming Cheng, Bo Li

We present ControlSR, a new method that can tame Diffusion Models for consistent real-world image super-resolution (Real-ISR). Previous Real-ISR models mostly focus on how to activate more generative priors of text-to-image diffusion models to make the output high-resolution (HR) images look better. However, since these methods rely too much on the generative priors, the content of the output images is often inconsistent with the input LR ones. To mitigate the above issue, in this work, we tame Diffusion Models by effectively utilizing LR information to impose stronger constraints on the control signals from ControlNet in the latent space. We show that our method can produce higher-quality control signals, which enables the super-resolution results to be more consistent with the LR image and leads to clearer visual results. In addition, we also propose an inference strategy that imposes constraints in the latent space using LR information, allowing for the simultaneous improvement of fidelity and generative ability. Experiments demonstrate that our model can achieve better performance across multiple metrics on several test sets and generate more consistent SR results with LR images than existing methods. Our code is available at https://github.com/HVision-NKU/ControlSR.

♻ ☆ StarGen: A Spatiotemporal Autoregression Framework with Video Diffusion Model for Scalable and Controllable Scene Generation

Shangjin Zhai, Zhichao Ye, Jialin Liu, Weijian Xie, Jiaqi Hu, Zhen Peng, Hua Xue, Danpeng Chen, Xiaomeng Wang, Lei Yang, Nan Wang, Haomin Liu, Guofeng Zhang

Recent advances in large reconstruction and generative models have significantly improved scene reconstruction and novel view generation. However, due to compute limitations, each inference with these large models is confined to a small area, making long-range consistent scene generation challenging. To address this, we propose StarGen, a novel framework that employs a pre-trained video diffusion model in an autoregressive manner for long-range scene generation. The generation of each video clip is conditioned on the 3D warping of spatially adjacent images and the temporally overlapping image from previously generated clips, improving spatiotemporal consistency in long-range scene generation with precise pose control. The spatiotemporal condition is compatible with various input conditions, facilitating diverse tasks, including sparse view interpolation, perpetual view generation, and layout-conditioned city generation. Quantitative and qualitative evaluations demonstrate StarGen's superior scalability, fidelity, and pose accuracy compared to state-of-the-art methods. Project page: https://zju3dv.github.io/StarGen.

♻ ☆ AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors ICLR 2025

Ruoxuan Feng, Jiangyu Hu, Wenke Xia, Tianci Gao, Ao Shen, Yuhao Sun, Bin Fang, Di Hu

Visuo-tactile sensors aim to emulate human tactile perception, enabling robots to precisely understand and manipulate objects. Over time, numerous meticulously designed visuo-tactile sensors have been integrated into robotic systems, aiding in completing various tasks. However, the distinct data characteristics of these low-standardized visuo-tactile sensors hinder the establishment of a powerful tactile perception system. We consider that the key to addressing this issue lies in learning unified multi-sensor representations, thereby integrating the sensors and promoting tactile knowledge transfer between them. To achieve unified representation of this nature, we introduce TacQuad, an aligned multi-modal multi-sensor tactile dataset from four different visuo-tactile sensors, which enables the explicit integration of various sensors. Recognizing that humans perceive the physical environment by acquiring diverse tactile information such as texture and pressure changes, we further propose to learn unified multi-sensor representations from both static and dynamic perspectives. By integrating tactile images and videos, we present AnyTouch, a unified static-dynamic multi-sensor representation learning framework with a multi-level structure, aimed at both enhancing comprehensive perceptual abilities and enabling effective cross-sensor transfer. This multi-level architecture captures pixel-level details from tactile data via masked modeling and enhances perception and transferability by learning semantic-level sensor-agnostic features through multi-modal alignment and cross-sensor matching. We provide a comprehensive analysis of multi-sensor transferability, and validate our method on various datasets and in the real-world pouring task. Experimental results show that our method outperforms existing methods, exhibits outstanding static and dynamic perception capabilities across various sensors.

comment: Accepted by ICLR 2025

♻ ☆ RainyGS: Efficient Rain Synthesis with Physically-Based Gaussian Splatting CVPR 2025

Qiyu Dai, Xingyu Ni, Qianfan Shen, Wenzheng Chen, Baoquan Chen, Mengyu Chu

We consider the problem of adding dynamic rain effects to in-the-wild scenes in a physically-correct manner. Recent advances in scene modeling have made significant progress, with NeRF and 3DGS techniques emerging as powerful tools for reconstructing complex scenes. However, while effective for novel view synthesis, these methods typically struggle with challenging scene editing tasks, such as physics-based rain simulation. In contrast, traditional physics-based simulations can generate realistic rain effects, such as raindrops and splashes, but they often rely on skilled artists to carefully set up high-fidelity scenes. This process lacks flexibility and scalability, limiting its applicability to broader, open-world environments. In this work, we introduce RainyGS, a novel approach that leverages the strengths of both physics-based modeling and 3DGS to generate photorealistic, dynamic rain effects in open-world scenes with physical accuracy. At the core of our method is the integration of physically-based raindrop and shallow water simulation techniques within the fast 3DGS rendering framework, enabling realistic and efficient simulations of raindrop behavior, splashes, and reflections. Our method supports synthesizing rain effects at over 30 fps, offering users flexible control over rain intensity -- from light drizzles to heavy downpours. We demonstrate that RainyGS performs effectively for both real-world outdoor scenes and large-scale driving scenarios, delivering more photorealistic and physically-accurate rain effects compared to state-of-the-art methods. Project page can be found at https://pku-vcl-geometry.github.io/RainyGS/

comment: CVPR 2025

♻ ☆ VFX Creator: Animated Visual Effect Generation with Controllable Diffusion Transformer

Xinyu Liu, Ailing Zeng, Wei Xue, Harry Yang, Wenhan Luo, Qifeng Liu, Yike Guo

Crafting magic and illusions is one of the most thrilling aspects of filmmaking, with visual effects (VFX) serving as the powerhouse behind unforgettable cinematic experiences. While recent advances in generative artificial intelligence have driven progress in generic image and video synthesis, the domain of controllable VFX generation remains relatively underexplored. In this work, we propose a novel paradigm for animated VFX generation as image animation, where dynamic effects are generated from user-friendly textual descriptions and static reference images. Our work makes two primary contributions: (i) Open-VFX, the first high-quality VFX video dataset spanning 15 diverse effect categories, annotated with textual descriptions, instance segmentation masks for spatial conditioning, and start-end timestamps for temporal control. (ii) VFX Creator, a simple yet effective controllable VFX generation framework based on a Video Diffusion Transformer. The model incorporates a spatial and temporal controllable LoRA adapter, requiring minimal training videos. Specifically, a plug-and-play mask control module enables instance-level spatial manipulation, while tokenized start-end motion timestamps embedded in the diffusion process, alongside the text encoder, allow precise temporal control over effect timing and pace. Extensive experiments on the Open-VFX test set demonstrate the superiority of the proposed system in generating realistic and dynamic effects, achieving state-of-the-art performance and generalization ability in both spatial and temporal controllability. Furthermore, we introduce a specialized metric to evaluate the precision of temporal control. By bridging traditional VFX techniques with generative approaches, VFX Creator unlocks new possibilities for efficient and high-quality video effect generation, making advanced VFX accessible to a broader audience.

♻ ☆ GaussianRoom: Improving 3D Gaussian Splatting with SDF Guidance and Monocular Cues for Indoor Scene Reconstruction

Haodong Xiang, Xinghui Li, Kai Cheng, Xiansong Lai, Wanting Zhang, Zhichao Liao, Long Zeng, Xueping Liu

Embodied intelligence requires precise reconstruction and rendering to simulate large-scale real-world data. Although 3D Gaussian Splatting (3DGS) has recently demonstrated high-quality results with real-time performance, it still faces challenges in indoor scenes with large, textureless regions, resulting in incomplete and noisy reconstructions due to poor point cloud initialization and underconstrained optimization. Inspired by the continuity of signed distance field (SDF), which naturally has advantages in modeling surfaces, we propose a unified optimization framework that integrates neural signed distance fields (SDFs) with 3DGS for accurate geometry reconstruction and real-time rendering. This framework incorporates a neural SDF field to guide the densification and pruning of Gaussians, enabling Gaussians to model scenes accurately even with poor initialized point clouds. Simultaneously, the geometry represented by Gaussians improves the efficiency of the SDF field by piloting its point sampling. Additionally, we introduce two regularization terms based on normal and edge priors to resolve geometric ambiguities in textureless areas and enhance detail accuracy. Extensive experiments in ScanNet and ScanNet++ show that our method achieves state-of-the-art performance in both surface reconstruction and novel view synthesis.

♻ ☆ Content-decoupled Contrastive Learning-based Implicit Degradation Modeling for Blind Image Super-Resolution

Jiang Yuan, Ji Ma, Bo Wang, Weiming Hu

Implicit degradation modeling-based blind super-resolution (SR) has attracted more increasing attention in the community due to its excellent generalization to complex degradation scenarios and wide application range. How to extract more discriminative degradation representations and fully adapt them to specific image features is the key to this task. In this paper, we propose a new Content-decoupled Contrastive Learning-based blind image super-resolution (CdCL) framework following the typical blind SR pipeline. This framework introduces negative-free contrastive learning technique for the first time to model the implicit degradation representation, in which a new cyclic shift sampling strategy is designed to ensure decoupling between content features and degradation features from the data perspective, thereby improving the purity and discriminability of the learned implicit degradation space. In addition, we propose a detail-aware implicit degradation adapting module that can better adapt degradation representations to specific LR features by enhancing the basic adaptation unit's perception of image details, significantly reducing the overall SR model complexity. Extensive experiments on synthetic and real data show that our method achieves highly competitive quantitative and qualitative results in various degradation settings while obviously reducing parameters and computational costs, validating the feasibility of designing practical and lightweight blind SR tools.

♻ ☆ Video-T1: Test-Time Scaling for Video Generation

Fangfu Liu, Hanyang Wang, Yimo Cai, Kaiyan Zhang, Xiaohang Zhan, Yueqi Duan

With the scale capability of increasing training data, model size, and computational cost, video generation has achieved impressive results in digital creation, enabling users to express creativity across various domains. Recently, researchers in Large Language Models (LLMs) have expanded the scaling to test-time, which can significantly improve LLM performance by using more inference-time computation. Instead of scaling up video foundation models through expensive training costs, we explore the power of Test-Time Scaling (TTS) in video generation, aiming to answer the question: if a video generation model is allowed to use non-trivial amount of inference-time compute, how much can it improve generation quality given a challenging text prompt. In this work, we reinterpret the test-time scaling of video generation as a searching problem to sample better trajectories from Gaussian noise space to the target video distribution. Specifically, we build the search space with test-time verifiers to provide feedback and heuristic algorithms to guide searching process. Given a text prompt, we first explore an intuitive linear search strategy by increasing noise candidates at inference time. As full-step denoising all frames simultaneously requires heavy test-time computation costs, we further design a more efficient TTS method for video generation called Tree-of-Frames (ToF) that adaptively expands and prunes video branches in an autoregressive manner. Extensive experiments on text-conditioned video generation benchmarks demonstrate that increasing test-time compute consistently leads to significant improvements in the quality of videos. Project page: https://liuff19.github.io/Video-T1

comment: Project page: https://liuff19.github.io/Video-T1

♻ ☆ Vision-Language Models for Acute Tuberculosis Diagnosis: A Multimodal Approach Combining Imaging and Clinical Data

Ananya Ganapthy, Praveen Shastry, Naveen Kumarasami, Anandakumar D, Keerthana R, Mounigasri M, Varshinipriya M, Kishore Prasath Venkatesh, Bargava Subramanian, Kalyan Sivasailam

Background: This study introduces a Vision-Language Model (VLM) leveraging SIGLIP and Gemma-3b architectures for automated acute tuberculosis (TB) screening. By integrating chest X-ray images and clinical notes, the model aims to enhance diagnostic accuracy and efficiency, particularly in resource-limited settings. Methods: The VLM combines visual data from chest X-rays with clinical context to generate detailed, context-aware diagnostic reports. The architecture employs SIGLIP for visual encoding and Gemma-3b for decoding, ensuring effective representation of acute TB-specific pathologies and clinical insights. Results: Key acute TB pathologies, including consolidation, cavities, and nodules, were detected with high precision (97percent) and recall (96percent). The model demonstrated strong spatial localization capabilities and robustness in distinguishing TB-positive cases, making it a reliable tool for acute TB diagnosis. Conclusion: The multimodal capability of the VLM reduces reliance on radiologists, providing a scalable solution for acute TB screening. Future work will focus on improving the detection of subtle pathologies and addressing dataset biases to enhance its generalizability and application in diverse global healthcare settings.

comment: 11 pages, 3 figures

♻ ☆ Generalizable Prompt Learning of CLIP: A Brief Overview

Fangming Cui, Yonggang Zhang, Xuan Wang, Xule Wang, Liang Xiao

Existing vision-language models (VLMs) such as CLIP have showcased an impressive capability to generalize well across various downstream tasks. These models leverage the synergy between visual and textual information, enabling them to understand and reason about the content present in images and text in a unified manner. This article provides a brief overview of CLIP based on few-shot prompt learning, including experimental data and technical characteristics of some methods. The purpose of this review is to provide a reference for researchers who have just started their research in generalizable prompting of CLIP through few-shot training for classification across 15 datasets and also to facilitate the integration of this field by researchers in other downstream tasks.

♻ ☆ ALLVB: All-in-One Long Video Understanding Benchmark AAAI 2025

Xichen Tan, Yuanjing Luo, Yunfan Ye, Fang Liu, Zhiping Cai

From image to video understanding, the capabilities of Multi-modal LLMs (MLLMs) are increasingly powerful. However, most existing video understanding benchmarks are relatively short, which makes them inadequate for effectively evaluating the long-sequence modeling capabilities of MLLMs. This highlights the urgent need for a comprehensive and integrated long video understanding benchmark to assess the ability of MLLMs thoroughly. To this end, we propose ALLVB (ALL-in-One Long Video Understanding Benchmark). ALLVB's main contributions include: 1) It integrates 9 major video understanding tasks. These tasks are converted into video QA formats, allowing a single benchmark to evaluate 9 different video understanding capabilities of MLLMs, highlighting the versatility, comprehensiveness, and challenging nature of ALLVB. 2) A fully automated annotation pipeline using GPT-4o is designed, requiring only human quality control, which facilitates the maintenance and expansion of the benchmark. 3) It contains 1,376 videos across 16 categories, averaging nearly 2 hours each, with a total of 252k QAs. To the best of our knowledge, it is the largest long video understanding benchmark in terms of the number of videos, average duration, and number of QAs. We have tested various mainstream MLLMs on ALLVB, and the results indicate that even the most advanced commercial models have significant room for improvement. This reflects the benchmark's challenging nature and demonstrates the substantial potential for development in long video understanding.

comment: AAAI 2025

♻ ☆ Zero-Shot Visual Concept Blending Without Text Guidance

Hiroya Makino, Takahiro Yamaguchi, Hiroyuki Sakai

We propose a novel, zero-shot image generation technique called "Visual Concept Blending" that provides fine-grained control over which features from multiple reference images are transferred to a source image. If only a single reference image is available, it is difficult to isolate which specific elements should be transferred. However, using multiple reference images, the proposed approach distinguishes between common and unique features by selectively incorporating them into a generated output. By operating within a partially disentangled Contrastive Language-Image Pre-training (CLIP) embedding space (from IP-Adapter), our method enables the flexible transfer of texture, shape, motion, style, and more abstract conceptual transformations without requiring additional training or text prompts. We demonstrate its effectiveness across a diverse range of tasks, including style transfer, form metamorphosis, and conceptual transformations, showing how subtle or abstract attributes (e.g., brushstroke style, aerodynamic lines, and dynamism) can be seamlessly combined into a new image. In a user study, participants accurately recognized which features were intended to be transferred. Its simplicity, flexibility, and high-level control make Visual Concept Blending valuable for creative fields such as art, design, and content creation, where combining specific visual qualities from multiple inspirations is crucial.

♻ ☆ Feed-Forward Bullet-Time Reconstruction of Dynamic Scenes from Monocular Videos

Hanxue Liang, Jiawei Ren, Ashkan Mirzaei, Antonio Torralba, Ziwei Liu, Igor Gilitschenski, Sanja Fidler, Cengiz Oztireli, Huan Ling, Zan Gojcic, Jiahui Huang

Recent advancements in static feed-forward scene reconstruction have demonstrated significant progress in high-quality novel view synthesis. However, these models often struggle with generalizability across diverse environments and fail to effectively handle dynamic content. We present BTimer (short for BulletTimer), the first motion-aware feed-forward model for real-time reconstruction and novel view synthesis of dynamic scenes. Our approach reconstructs the full scene in a 3D Gaussian Splatting representation at a given target ('bullet') timestamp by aggregating information from all the context frames. Such a formulation allows BTimer to gain scalability and generalization by leveraging both static and dynamic scene datasets. Given a casual monocular dynamic video, BTimer reconstructs a bullet-time scene within 150ms while reaching state-of-the-art performance on both static and dynamic scene datasets, even compared with optimization-based approaches.

comment: Project website: https://research.nvidia.com/labs/toronto-ai/bullet-timer/

♻ ☆ Diffusion Models in 3D Vision: A Survey

Zhen Wang, Dongyuan Li, Yaozu Wu, Tianyu He, Jiang Bian, Renhe Jiang

In recent years, 3D vision has become a crucial field within computer vision, powering a wide range of applications such as autonomous driving, robotics, augmented reality, and medical imaging. This field relies on accurate perception, understanding, and reconstruction of 3D scenes from 2D images or text data sources. Diffusion models, originally designed for 2D generative tasks, offer the potential for more flexible, probabilistic methods that can better capture the variability and uncertainty present in real-world 3D data. In this paper, we review the state-of-the-art methods that use diffusion models for 3D visual tasks, including but not limited to 3D object generation, shape completion, point-cloud reconstruction, and scene construction. We provide an in-depth discussion of the underlying mathematical principles of diffusion models, outlining their forward and reverse processes, as well as the various architectural advancements that enable these models to work with 3D datasets. We also discuss the key challenges in applying diffusion models to 3D vision, such as handling occlusions and varying point densities, and the computational demands of high-dimensional data. Finally, we discuss potential solutions, including improving computational efficiency, enhancing multimodal fusion, and exploring the use of large-scale pretraining for better generalization across 3D tasks. This paper serves as a foundation for future exploration and development in this rapidly evolving field.

♻ ☆ Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysis CVPR 2025

Jiangyong Huang, Baoxiong Jia, Yan Wang, Ziyu Zhu, Xiongkun Linghu, Qing Li, Song-Chun Zhu, Siyuan Huang

Existing 3D vision-language (3D-VL) benchmarks fall short in evaluating 3D-VL models, creating a "mist" that obscures rigorous insights into model capabilities and 3D-VL tasks. This mist persists due to three key limitations. First, flawed test data, like ambiguous referential text in the grounding task, can yield incorrect and unreliable test results. Second, oversimplified metrics such as simply averaging accuracy per question answering (QA) pair, cannot reveal true model capability due to their vulnerability to language variations. Third, existing benchmarks isolate the grounding and QA tasks, disregarding the underlying coherence that QA should be based on solid grounding capabilities. To unveil the "mist", we propose Beacon3D, a benchmark for 3D-VL grounding and QA tasks, delivering a perspective shift in the evaluation of 3D-VL understanding. Beacon3D features (i) high-quality test data with precise and natural language, (ii) object-centric evaluation with multiple tests per object to ensure robustness, and (iii) a novel chain-of-analysis paradigm to address language robustness and model performance coherence across grounding and QA. Our evaluation of state-of-the-art 3D-VL models on Beacon3D reveals that (i) object-centric evaluation elicits true model performance and particularly weak generalization in QA; (ii) grounding-QA coherence remains fragile in current 3D-VL models, and (iii) incorporating large language models (LLMs) to 3D-VL models, though as a prevalent practice, hinders grounding capabilities and has yet to elevate QA capabilities. We hope Beacon3D and our comprehensive analysis could benefit the 3D-VL community towards faithful developments.

comment: CVPR 2025. Project page: https://beacon-3d.github.io

♻ ☆ MagicPose4D: Crafting Articulated Models with Appearance and Motion Control

Hao Zhang, Di Chang, Fang Li, Mohammad Soleymani, Narendra Ahuja

With the success of 2D and 3D visual generative models, there is growing interest in generating 4D content. Existing methods primarily rely on text prompts to produce 4D content, but they often fall short of accurately defining complex or rare motions. To address this limitation, we propose MagicPose4D, a novel framework for refined control over both appearance and motion in 4D generation. Unlike current 4D generation methods, MagicPose4D accepts monocular videos or mesh sequences as motion prompts, enabling precise and customizable motion control. MagicPose4D comprises two key modules: (i) Dual-Phase 4D Reconstruction Module, which operates in two phases. The first phase focuses on capturing the model's shape using accurate 2D supervision and less accurate but geometrically informative 3D pseudo-supervision without imposing skeleton constraints. The second phase extracts the 3D motion (skeleton poses) using more accurate pseudo-3D supervision, obtained in the first phase and introduces kinematic chain-based skeleton constraints to ensure physical plausibility. Additionally, we propose a Global-local Chamfer loss that aligns the overall distribution of predicted mesh vertices with the supervision while maintaining part-level alignment without extra annotations. (ii) Cross-category Motion Transfer Module, which leverages the extracted motion from the 4D reconstruction module and uses a kinematic-chain-based skeleton to achieve cross-category motion transfer. It ensures smooth transitions between frames through dynamic rigidity, facilitating robust generalization without additional training. Through extensive experiments, we demonstrate that MagicPose4D significantly improves the accuracy and consistency of 4D content generation, outperforming existing methods in various benchmarks.

comment: Project Page: https://magicpose4d.github.io/

♻ ☆ Phase-shifted remote photoplethysmography for estimating heart rate and blood pressure from facial video

Gyutae Hwang, Sang Jun Lee

Human health can be critically affected by cardiovascular diseases, such as hypertension, arrhythmias, and stroke. Heart rate and blood pressure are important biometric information for the monitoring of cardiovascular system and early diagnosis of cardiovascular diseases. Existing methods for estimating the heart rate are based on electrocardiography and photoplethyomography, which require contacting the sensor to the skin surface. Moreover, catheter and cuff-based methods for measuring blood pressure cause inconvenience and have limited applicability. Therefore, in this thesis, we propose a vision-based method for estimating the heart rate and blood pressure. This thesis proposes a 2-stage deep learning framework consisting of a dual remote photoplethysmography network (DRP-Net) and bounded blood pressure network (BBP-Net). In the first stage, DRP-Net infers remote photoplethysmography (rPPG) signals for the acral and facial regions, and these phase-shifted rPPG signals are utilized to estimate the heart rate. In the second stage, BBP-Net integrates temporal features and analyzes phase discrepancy between the acral and facial rPPG signals to estimate SBP and DBP values. To improve the accuracy of estimating the heart rate, we employed a data augmentation method based on a frame interpolation model. Moreover, we designed BBP-Net to infer blood pressure within a predefined range by incorporating a scaled sigmoid function. Our method resulted in estimating the heart rate with the mean absolute error (MAE) of 1.78 BPM, reducing the MAE by 34.31 % compared to the recent method, on the MMSE-HR dataset. The MAE for estimating the systolic blood pressure (SBP) and diastolic blood pressure (DBP) were 10.19 mmHg and 7.09 mmHg. On the V4V dataset, the MAE for the heart rate, SBP, and DBP were 3.83 BPM, 13.64 mmHg, and 9.4 mmHg, respectively.

comment: 13 pages, 10 figures

♻ ☆ VRM: Knowledge Distillation via Virtual Relation Matching

Weijia Zhang, Fei Xie, Weidong Cai, Chao Ma

Knowledge distillation (KD) aims to transfer the knowledge of a more capable yet cumbersome teacher model to a lightweight student model. In recent years, relation-based KD methods have fallen behind, as their instance-matching counterparts dominate in performance. In this paper, we revive relational KD by identifying and tackling several key issues in relation-based methods, including their susceptibility to overfitting and spurious responses. Specifically, we transfer novelly constructed affinity graphs that compactly encapsulate a wealth of beneficial inter-sample, inter-class, and inter-view correlations by exploiting virtual views and relations as a new kind of knowledge. As a result, the student has access to richer guidance signals and stronger regularisation throughout the distillation process. To further mitigate the adverse impact of spurious responses, we prune the affinity graphs by dynamically detaching redundant and unreliable edges. Extensive experiments on CIFAR-100 and ImageNet datasets demonstrate the superior performance of the proposed virtual relation matching (VRM) method over a range of models, architectures, and set-ups. For instance, VRM for the first time hits 74.0% accuracy for ResNet50-to-MobileNetV2 distillation on ImageNet, and improves DeiT-T by 14.44% on CIFAR-100 with a ResNet56 teacher. Thorough analyses are also conducted to gauge the soundness, properties, and complexity of our designs. Code and models will be released.

♻ ☆ Controllable Human Image Generation with Personalized Multi-Garments CVPR 2025

Yisol Choi, Sangkyung Kwak, Sihyun Yu, Hyungwon Choi, Jinwoo Shin

We present BootComp, a novel framework based on text-to-image diffusion models for controllable human image generation with multiple reference garments. Here, the main bottleneck is data acquisition for training: collecting a large-scale dataset of high-quality reference garment images per human subject is quite challenging, i.e., ideally, one needs to manually gather every single garment photograph worn by each human. To address this, we propose a data generation pipeline to construct a large synthetic dataset, consisting of human and multiple-garment pairs, by introducing a model to extract any reference garment images from each human image. To ensure data quality, we also propose a filtering strategy to remove undesirable generated data based on measuring perceptual similarities between the garment presented in human image and extracted garment. Finally, by utilizing the constructed synthetic dataset, we train a diffusion model having two parallel denoising paths that use multiple garment images as conditions to generate human images while preserving their fine-grained details. We further show the wide-applicability of our framework by adapting it to different types of reference-based generation in the fashion domain, including virtual try-on, and controllable human image generation with other conditions, e.g., pose, face, etc.

comment: CVPR 2025. Project page: https://omnious.github.io/BootComp

♻ ☆ VisRL: Intention-Driven Visual Perception via Reinforced Reasoning

Zhangquan Chen, Xufang Luo, Dongsheng Li

Visual understanding is inherently intention-driven - humans selectively focus on different regions of a scene based on their goals. Recent advances in large multimodal models (LMMs) enable flexible expression of such intentions through natural language, allowing queries to guide visual reasoning processes. Frameworks like Visual Chain-of-Thought have demonstrated the benefit of incorporating explicit reasoning steps, where the model predicts a focus region before answering a query. However, existing approaches rely heavily on supervised training with annotated intermediate bounding boxes, which severely limits scalability due to the combinatorial explosion of intention-region pairs. To overcome this limitation, we propose VisRL, the first framework that applies reinforcement learning (RL) to the problem of intention-driven visual perception. VisRL optimizes the entire visual reasoning process using only reward signals. By treating intermediate focus selection as an internal decision optimized through trial-and-error, our method eliminates the need for costly region annotations while aligning more closely with how humans learn to perceive the world. Extensive experiments across multiple benchmarks show that VisRL consistently outperforms strong baselines, demonstrating both its effectiveness and its strong generalization across different LMMs. Our code is available at https://github.com/zhangquanchen/VisRL.

comment: 18pages,11 figures

♻ ☆ Diffusion State-Guided Projected Gradient for Inverse Problems ICLR 2025

Rayhan Zirvi, Bahareh Tolooshams, Anima Anandkumar

Recent advancements in diffusion models have been effective in learning data priors for solving inverse problems. They leverage diffusion sampling steps for inducing a data prior while using a measurement guidance gradient at each step to impose data consistency. For general inverse problems, approximations are needed when an unconditionally trained diffusion model is used since the measurement likelihood is intractable, leading to inaccurate posterior sampling. In other words, due to their approximations, these methods fail to preserve the generation process on the data manifold defined by the diffusion prior, leading to artifacts in applications such as image restoration. To enhance the performance and robustness of diffusion models in solving inverse problems, we propose Diffusion State-Guided Projected Gradient (DiffStateGrad), which projects the measurement gradient onto a subspace that is a low-rank approximation of an intermediate state of the diffusion process. DiffStateGrad, as a module, can be added to a wide range of diffusion-based inverse solvers to improve the preservation of the diffusion process on the prior manifold and filter out artifact-inducing components. We highlight that DiffStateGrad improves the robustness of diffusion models in terms of the choice of measurement guidance step size and noise while improving the worst-case performance. Finally, we demonstrate that DiffStateGrad improves upon the state-of-the-art on linear and nonlinear image restoration inverse problems. Our code is available at https://github.com/Anima-Lab/DiffStateGrad.

comment: Published as a conference paper at ICLR 2025. RZ and BT have equal contributions

♻ ☆ Retrieval-augmented Few-shot Medical Image Segmentation with Foundation Models

Lin Zhao, Xiao Chen, Eric Z. Chen, Yikang Liu, Terrence Chen, Shanhui Sun

Medical image segmentation is crucial for clinical decision-making, but the scarcity of annotated data presents significant challenges. Few-shot segmentation (FSS) methods show promise but often require training on the target domain and struggle to generalize across different modalities. Similarly, adapting foundation models like the Segment Anything Model (SAM) for medical imaging has limitations, including the need for finetuning and domain-specific adaptation. To address these issues, we propose a novel method that adapts DINOv2 and Segment Anything Model 2 (SAM 2) for retrieval-augmented few-shot medical image segmentation. Our approach uses DINOv2's feature as query to retrieve similar samples from limited annotated data, which are then encoded as memories and stored in memory bank. With the memory attention mechanism of SAM 2, the model leverages these memories as conditions to generate accurate segmentation of the target image. We evaluated our framework on three medical image segmentation tasks, demonstrating superior performance and generalizability across various modalities without the need for any retraining or finetuning. Overall, this method offers a practical and effective solution for few-shot medical image segmentation and holds significant potential as a valuable annotation tool in clinical applications.

♻ ☆ VideoMind: A Chain-of-LoRA Agent for Long Video Reasoning

Ye Liu, Kevin Qinghong Lin, Chang Wen Chen, Mike Zheng Shou

Videos, with their unique temporal dimension, demand precise grounded understanding, where answers are directly linked to visual, interpretable evidence. Despite significant breakthroughs in reasoning capabilities within Large Language Models, multi-modal reasoning - especially for videos - remains unexplored. In this work, we introduce VideoMind, a novel video-language agent designed for temporal-grounded video understanding. VideoMind incorporates two key innovations: (i) We identify essential capabilities for video temporal reasoning and develop a role-based agentic workflow, including a planner for coordinating different roles, a grounder for temporal localization, a verifier to assess temporal interval accuracy, and an answerer for question-answering. (ii) To efficiently integrate these diverse roles, we propose a novel Chain-of-LoRA strategy, enabling seamless role-switching via lightweight LoRA adaptors while avoiding the overhead of multiple models, thus balancing efficiency and flexibility. Extensive experiments on 14 public benchmarks, including 3 on grounded video question-answering (Grounded VideoQA), 6 on video temporal grounding (VTG), and 5 on general video question-answering (VideoQA), verify that our agent achieves state-of-the-art performance on diverse video understanding tasks, underscoring its effectiveness in advancing video agent and long-form temporal reasoning.

comment: Project Page: https://videomind.github.io/

♻ ☆ Without Paired Labeled Data: An End-to-End Self-Supervised Paradigm for UAV-View Geo-Localization

Zhongwei Chen, Zhao-Xu Yang, Hai-Jun Rong

UAV-View Geo-Localization (UVGL) aims to achieve accurate localization of unmanned aerial vehicles (UAVs) by retrieving the most relevant GPS-tagged satellite images. However, existing methods heavily rely on pre-paired UAV-satellite images for supervised learning. Such dependency not only incurs high annotation costs but also severely limits scalability and practical deployment in open-world UVGL scenarios. To address these limitations, we propose an end-to-end self-supervised UVGL method. Our method leverages a shallow backbone network to extract initial features, employs clustering to generate pseudo labels, and adopts a dual-path contrastive learning architecture to learn discriminative intra-view representations. Furthermore, our method incorporates two core modules, the dynamic hierarchical memory learning module and the information consistency evolution learning module. The dynamic hierarchical memory learning module combines short-term and long-term memory to enhance intra-view feature consistency and discriminability. Meanwhile, the information consistency evolution learning module leverages a neighborhood-driven dynamic constraint mechanism to systematically capture implicit cross-view semantic correlations, thereby improving cross-view feature alignment. To further stabilize and strengthen the self-supervised training process, a pseudo-label enhancement strategy is introduced, which refines the quality of pseudo supervision. Our method ultimately constructs a unified cross-view feature representation space under self-supervised settings. Extensive experiments on three public benchmark datasets demonstrate that the proposed method consistently outperforms existing self-supervised methods and even surpasses several state-of-the-art supervised methods. Our code is available at https://github.com/ISChenawei/DMNIL.

♻ ☆ HumanDreamer: Generating Controllable Human-Motion Videos via Decoupled Generation

Boyuan Wang, Xiaofeng Wang, Chaojun Ni, Guosheng Zhao, Zhiqin Yang, Zheng Zhu, Muyang Zhang, Yukun Zhou, Xinze Chen, Guan Huang, Lihong Liu, Xingang Wang

Human-motion video generation has been a challenging task, primarily due to the difficulty inherent in learning human body movements. While some approaches have attempted to drive human-centric video generation explicitly through pose control, these methods typically rely on poses derived from existing videos, thereby lacking flexibility. To address this, we propose HumanDreamer, a decoupled human video generation framework that first generates diverse poses from text prompts and then leverages these poses to generate human-motion videos. Specifically, we propose MotionVid, the largest dataset for human-motion pose generation. Based on the dataset, we present MotionDiT, which is trained to generate structured human-motion poses from text prompts. Besides, a novel LAMA loss is introduced, which together contribute to a significant improvement in FID by 62.4%, along with respective enhancements in R-precision for top1, top2, and top3 by 41.8%, 26.3%, and 18.3%, thereby advancing both the Text-to-Pose control accuracy and FID metrics. Our experiments across various Pose-to-Video baselines demonstrate that the poses generated by our method can produce diverse and high-quality human-motion videos. Furthermore, our model can facilitate other downstream tasks, such as pose sequence prediction and 2D-3D motion lifting.

comment: Project Page: https://humandreamer.github.io

♻ ☆ Data-Free Group-Wise Fully Quantized Winograd Convolution via Learnable Scales CVPR 2025

Shuokai Pan, Gerti Tuzi, Sudarshan Sreeram, Dibakar Gope

Despite the revolutionary breakthroughs of large-scale text-to-image diffusion models for complex vision and downstream tasks, their extremely high computational and storage costs limit their usability. Quantization of diffusion models has been explored in recent works to reduce compute costs and memory bandwidth usage. To further improve inference time, fast convolution algorithms such as Winograd can be used for convolution layers, which account for a significant portion of computations in diffusion models. However, the significant quality loss of fully quantized Winograd using existing coarser-grained post-training quantization methods, combined with the complexity and cost of finetuning the Winograd transformation matrices for such large models to recover quality, makes them unsuitable for large-scale foundation models. Motivated by the presence of a large range of values in them, we investigate the impact of finer-grained group-wise quantization in quantizing diffusion models. While group-wise quantization can largely handle the fully quantized Winograd convolution, it struggles to deal with the large distribution imbalance in a sizable portion of the Winograd domain computation. To reduce range differences in the Winograd domain, we propose finetuning only the scale parameters of the Winograd transform matrices without using any domain-specific training data. Because our method does not depend on any training data, the generalization performance of quantized diffusion models is safely guaranteed. For text-to-image generation task, the 8-bit fully-quantized diffusion model with Winograd provides near-lossless quality (FID and CLIP scores) in comparison to the full-precision model. For image classification, our method outperforms the state-of-the-art Winograd PTQ method by 1.62% and 2.56% in top-1 ImageNet accuracy on ResNet18 and ResNet-34, respectively, with Winograd F(6, 3).

comment: Accepted by CVPR 2025

♻ ☆ Visual Acoustic Fields

Yuelei Li, Hyunjin Kim, Fangneng Zhan, Ri-Zhao Qiu, Mazeyu Ji, Xiaojun Shan, Xueyan Zou, Paul Liang, Hanspeter Pfister, Xiaolong Wang

Objects produce different sounds when hit, and humans can intuitively infer how an object might sound based on its appearance and material properties. Inspired by this intuition, we propose Visual Acoustic Fields, a framework that bridges hitting sounds and visual signals within a 3D space using 3D Gaussian Splatting (3DGS). Our approach features two key modules: sound generation and sound localization. The sound generation module leverages a conditional diffusion model, which takes multiscale features rendered from a feature-augmented 3DGS to generate realistic hitting sounds. Meanwhile, the sound localization module enables querying the 3D scene, represented by the feature-augmented 3DGS, to localize hitting positions based on the sound sources. To support this framework, we introduce a novel pipeline for collecting scene-level visual-sound sample pairs, achieving alignment between captured images, impact locations, and corresponding sounds. To the best of our knowledge, this is the first dataset to connect visual and acoustic signals in a 3D context. Extensive experiments on our dataset demonstrate the effectiveness of Visual Acoustic Fields in generating plausible impact sounds and accurately localizing impact sources. Our project page is at https://yuelei0428.github.io/projects/Visual-Acoustic-Fields/.

♻ ☆ Astrea: A MOE-based Visual Understanding Model with Progressive Alignment

Xiaoda Yang, JunYu Lu, Hongshun Qiu, Sijing Li, Hao Li, Shengpeng Ji, Xudong Tang, Jiayang Xu, Jiaqi Duan, Ziyue Jiang, Cong Lin, Sihang Cai, Zejian Xie, Zhuoyang Song, Songxin Zhang

Vision-Language Models (VLMs) based on Mixture-of-Experts (MoE) architectures have emerged as a pivotal paradigm in multimodal understanding, offering a powerful framework for integrating visual and linguistic information. However, the increasing complexity and diversity of tasks present significant challenges in coordinating load balancing across heterogeneous visual experts, where optimizing one specialist's performance often compromises others' capabilities. To address task heterogeneity and expert load imbalance, we propose Astrea, a novel multi-expert collaborative VLM architecture based on progressive pre-alignment. Astrea introduces three key innovations: 1) A heterogeneous expert coordination mechanism that integrates four specialized models (detection, segmentation, classification, captioning) into a comprehensive expert matrix covering essential visual comprehension elements; 2) A dynamic knowledge fusion strategy featuring progressive pre-alignment to harmonize experts within the VLM latent space through contrastive learning, complemented by probabilistically activated stochastic residual connections to preserve knowledge continuity; 3) An enhanced optimization framework utilizing momentum contrastive learning for long-range dependency modeling and adaptive weight allocators for real-time expert contribution calibration. Extensive evaluations across 12 benchmark tasks spanning VQA, image captioning, and cross-modal retrieval demonstrate Astrea's superiority over state-of-the-art models, achieving an average performance gain of +4.7\%. This study provides the first empirical demonstration that progressive pre-alignment strategies enable VLMs to overcome task heterogeneity limitations, establishing new methodological foundations for developing general-purpose multimodal agents.

♻ ☆ 4D LangSplat: 4D Language Gaussian Splatting via Multimodal Large Language Models CVPR 2025

Wanhua Li, Renping Zhou, Jiawei Zhou, Yingwei Song, Johannes Herter, Minghan Qin, Gao Huang, Hanspeter Pfister

Learning 4D language fields to enable time-sensitive, open-ended language queries in dynamic scenes is essential for many real-world applications. While LangSplat successfully grounds CLIP features into 3D Gaussian representations, achieving precision and efficiency in 3D static scenes, it lacks the ability to handle dynamic 4D fields as CLIP, designed for static image-text tasks, cannot capture temporal dynamics in videos. Real-world environments are inherently dynamic, with object semantics evolving over time. Building a precise 4D language field necessitates obtaining pixel-aligned, object-wise video features, which current vision models struggle to achieve. To address these challenges, we propose 4D LangSplat, which learns 4D language fields to handle time-agnostic or time-sensitive open-vocabulary queries in dynamic scenes efficiently. 4D LangSplat bypasses learning the language field from vision features and instead learns directly from text generated from object-wise video captions via Multimodal Large Language Models (MLLMs). Specifically, we propose a multimodal object-wise video prompting method, consisting of visual and text prompts that guide MLLMs to generate detailed, temporally consistent, high-quality captions for objects throughout a video. These captions are encoded using a Large Language Model into high-quality sentence embeddings, which then serve as pixel-aligned, object-specific feature supervision, facilitating open-vocabulary text queries through shared embedding spaces. Recognizing that objects in 4D scenes exhibit smooth transitions across states, we further propose a status deformable network to model these continuous changes over time effectively. Our results across multiple benchmarks demonstrate that 4D LangSplat attains precise and efficient results for both time-sensitive and time-agnostic open-vocabulary queries.

comment: CVPR 2025. Project Page: https://4d-langsplat.github.io

♻ ☆ Learned Image Compression and Restoration for Digital Pathology

SeonYeong Lee, EonSeung Seong, DongEon Lee, SiYeoul Lee, Yubin Cho, Chunsu Park, Seonho Kim, MinKyung Seo, YoungSin Ko, MinWoo Kim

Digital pathology images play a crucial role in medical diagnostics, but their ultra-high resolution and large file sizes pose significant challenges for storage, transmission, and real-time visualization. To address these issues, we propose CLERIC, a novel deep learning-based image compression framework designed specifically for whole slide images (WSIs). CLERIC integrates a learnable lifting scheme and advanced convolutional techniques to enhance compression efficiency while preserving critical pathological details. Our framework employs a lifting-scheme transform in the analysis stage to decompose images into low- and high-frequency components, enabling more structured latent representations. These components are processed through parallel encoders incorporating Deformable Residual Blocks (DRB) and Recurrent Residual Blocks (R2B) to improve feature extraction and spatial adaptability. The synthesis stage applies an inverse lifting transform for effective image reconstruction, ensuring high-fidelity restoration of fine-grained tissue structures. We evaluate CLERIC on a digital pathology image dataset and compare its performance against state-of-the-art learned image compression (LIC) models. Experimental results demonstrate that CLERIC achieves superior rate-distortion (RD) performance, significantly reducing storage requirements while maintaining high diagnostic image quality. Our study highlights the potential of deep learning-based compression in digital pathology, facilitating efficient data management and long-term storage while ensuring seamless integration into clinical workflows and AI-assisted diagnostic systems. Code and models are available at: https://github.com/pnu-amilab/CLERIC.

♻ ☆ TextCrafter: Accurately Rendering Multiple Texts in Complex Visual Scenes

Nikai Du, Zhennan Chen, Zhizhou Chen, Shan Gao, Xi Chen, Zhengkai Jiang, Jian Yang, Ying Tai

This paper explores the task of Complex Visual Text Generation (CVTG), which centers on generating intricate textual content distributed across diverse regions within visual images. In CVTG, image generation models often rendering distorted and blurred visual text or missing some visual text. To tackle these challenges, we propose TextCrafter, a novel multi-visual text rendering method. TextCrafter employs a progressive strategy to decompose complex visual text into distinct components while ensuring robust alignment between textual content and its visual carrier. Additionally, it incorporates a token focus enhancement mechanism to amplify the prominence of visual text during the generation process. TextCrafter effectively addresses key challenges in CVTG tasks, such as text confusion, omissions, and blurriness. Moreover, we present a new benchmark dataset, CVTG-2K, tailored to rigorously evaluate the performance of generative models on CVTG tasks. Extensive experiments demonstrate that our method surpasses state-of-the-art approaches.

♻ ☆ EventMamba: Enhancing Spatio-Temporal Locality with State Space Models for Event-Based Video Reconstruction

Chengjie Ge, Xueyang Fu, Peng He, Kunyu Wang, Chengzhi Cao, Zheng-Jun Zha

Leveraging its robust linear global modeling capability, Mamba has notably excelled in computer vision. Despite its success, existing Mamba-based vision models have overlooked the nuances of event-driven tasks, especially in video reconstruction. Event-based video reconstruction (EBVR) demands spatial translation invariance and close attention to local event relationships in the spatio-temporal domain. Unfortunately, conventional Mamba algorithms apply static window partitions and standard reshape scanning methods, leading to significant losses in local connectivity. To overcome these limitations, we introduce EventMamba--a specialized model designed for EBVR tasks. EventMamba innovates by incorporating random window offset (RWO) in the spatial domain, moving away from the restrictive fixed partitioning. Additionally, it features a new consistent traversal serialization approach in the spatio-temporal domain, which maintains the proximity of adjacent events both spatially and temporally. These enhancements enable EventMamba to retain Mamba's robust modeling capabilities while significantly preserving the spatio-temporal locality of event data. Comprehensive testing on multiple datasets shows that EventMamba markedly enhances video reconstruction, drastically improving computation speed while delivering superior visual quality compared to Transformer-based methods.

♻ ☆ Where am I? Cross-View Geo-localization with Natural Language Descriptions

Junyan Ye, Honglin Lin, Leyan Ou, Dairong Chen, Zihao Wang, Qi Zhu, Conghui He, Weijia Li

Cross-view geo-localization identifies the locations of street-view images by matching them with geo-tagged satellite images or OSM. However, most existing studies focus on image-to-image retrieval, with fewer addressing text-guided retrieval, a task vital for applications like pedestrian navigation and emergency response. In this work, we introduce a novel task for cross-view geo-localization with natural language descriptions, which aims to retrieve corresponding satellite images or OSM database based on scene text descriptions. To support this task, we construct the CVG-Text dataset by collecting cross-view data from multiple cities and employing a scene text generation approach that leverages the annotation capabilities of Large Multimodal Models to produce high-quality scene text descriptions with localization details. Additionally, we propose a novel text-based retrieval localization method, CrossText2Loc, which improves recall by 10% and demonstrates excellent long-text retrieval capabilities. In terms of explainability, it not only provides similarity scores but also offers retrieval reasons. More information can be found at https://yejy53.github.io/CVG-Text/ .

comment: 11 pages, 6 figures

♻ ☆ Alleviating Hallucinations in Large Vision-Language Models through Hallucination-Induced Optimization NeurIPS 2024

Xinyu Lyu, Beitao Chen, Lianli Gao, Jingkuan Song, Heng Tao Shen

Although Large Visual Language Models (LVLMs) have demonstrated exceptional abilities in understanding multimodal data, they invariably suffer from hallucinations, leading to a disconnect between the generated text and the corresponding images. Almost all current visual contrastive decoding methods attempt to mitigate these hallucinations by introducing visual uncertainty information that appropriately widens the contrastive logits gap between hallucinatory and targeted ones. However, due to uncontrollable nature of the global visual uncertainty, they struggle to precisely induce the hallucinatory tokens, which severely limits their effectiveness in mitigating hallucinations and may even lead to the generation of undesired hallucinations. To tackle this issue, we conducted the theoretical analysis to promote the effectiveness of contrast decoding. Building on this insight, we introduce a novel optimization strategy named Hallucination-Induced Optimization (HIO). This strategy seeks to amplify the contrast between hallucinatory and targeted tokens relying on a fine-tuned theoretical preference model (i.e., Contrary Bradley-Terry Model), thereby facilitating efficient contrast decoding to alleviate hallucinations in LVLMs. Extensive experimental research demonstrates that our HIO strategy can effectively reduce hallucinations in LVLMs, outperforming state-of-the-art methods across various benchmarks.

comment: Accepted by NeurIPS 2024

♻ ☆ On-device Sora: Enabling Training-Free Diffusion-based Text-to-Video Generation for Mobile Devices

Bosung Kim, Kyuhwan Lee, Isu Jeong, Jungmin Cheon, Yeojin Lee, Seulki Lee

We present On-device Sora, the first model training-free solution for diffusion-based on-device text-to-video generation that operates efficiently on smartphone-grade devices. To address the challenges of diffusion-based text-to-video generation on computation- and memory-limited mobile devices, the proposed On-device Sora applies three novel techniques to pre-trained video generative models. First, Linear Proportional Leap (LPL) reduces the excessive denoising steps required in video diffusion through an efficient leap-based approach. Second, Temporal Dimension Token Merging (TDTM) minimizes intensive token-processing computation in attention layers by merging consecutive tokens along the temporal dimension. Third, Concurrent Inference with Dynamic Loading (CI-DL) dynamically partitions large models into smaller blocks and loads them into memory for concurrent model inference, effectively addressing the challenges of limited device memory. We implement On-device Sora on the iPhone 15 Pro, and the experimental evaluations show that it is capable of generating high-quality videos on the device, comparable to those produced by high-end GPUs. These results show that On-device Sora enables efficient and high-quality video generation on resource-constrained mobile devices. We envision the proposed On-device Sora as a significant first step toward democratizing state-of-the-art generative technologies, enabling video generation on commodity mobile and embedded devices without resource-intensive re-training for model optimization (compression). The code implementation is available at a GitHub repository(https://github.com/eai-lab/On-device-Sora).

comment: Replicated Submission. arXiv:2502.04363 submitted as second version of the paper

♻ ☆ A novel algorithm for optimizing bundle adjustment in image sequence alignment

Hailin Xu, Hongxia Wang, Huanshui Zhang

The Bundle Adjustment (BA) model is commonly optimized using a nonlinear least squares method, with the Levenberg-Marquardt (L-M) algorithm being a typical choice. However, despite the L-M algorithm's effectiveness, its sensitivity to initial conditions often results in slower convergence when applied to poorly conditioned datasets, motivating the exploration of alternative optimization strategies. This paper introduces a novel algorithm for optimizing the BA model in the context of image sequence alignment for cryo-electron tomography, utilizing optimal control theory to directly optimize general nonlinear functions. The proposed Optimal Control Algorithm (OCA) exhibits superior convergence rates and effectively mitigates the oscillatory behavior frequently observed in L-M algorithm. Extensive experiments on both synthetic and real-world datasets were conducted to evaluate the algorithm's performance. The results demonstrate that the OCA achieves faster convergence compared to the L-M algorithm. Moreover, the incorporation of a bisection-based update procedure significantly enhances the OCA's performance, particularly in poorly initialized datasets. These findings indicate that the OCA can substantially improve the efficiency of 3D reconstructions in cryo-electron tomography.

♻ ☆ WaveFormer: A 3D Transformer with Wavelet-Driven Feature Representation for Efficient Medical Image Segmentation

Md Mahfuz Al Hasan, Mahdi Zaman, Abdul Jawad, Alberto Santamaria-Pang, Ho Hin Lee, Ivan Tarapov, Kyle See, Md Shah Imran, Antika Roy, Yaser Pourmohammadi Fallah, Navid Asadizanjani, Reza Forghani

Transformer-based architectures have advanced medical image analysis by effectively modeling long-range dependencies, yet they often struggle in 3D settings due to substantial memory overhead and insufficient capture of fine-grained local features. We address these limitations with WaveFormer, a novel 3D-transformer that: i) leverages the fundamental frequency-domain properties of features for contextual representation, and ii) is inspired by the top-down mechanism of the human visual recognition system, making it a biologically motivated architecture. By employing discrete wavelet transformations (DWT) at multiple scales, WaveFormer preserves both global context and high-frequency details while replacing heavy upsampling layers with efficient wavelet-based summarization and reconstruction. This significantly reduces the number of parameters, which is critical for real-world deployment where computational resources and training times are constrained. Furthermore, the model is generic and easily adaptable to diverse applications. Evaluations on BraTS2023, FLARE2021, and KiTS2023 demonstrate performance on par with state-of-the-art methods while offering substantially lower computational complexity.

♻ ☆ A Comparative Tutorial of the Histogram-based Image Segmentation Methods

ZhenZhou Wang

The histogram of an image is the accurate graphical representation of the numerical grayscale distribution and it is also an estimate of the probability distribution of image pixels. Therefore, histogram has been widely adopted to calculate the clustering means and partitioning thresholds for image segmentation. There have been many classical histogram-based image segmentation methods proposed and played important roles in both academics and industry. In this tutorial, the histories and recent advances of the histogram-based image segmentation techniques are first reviewed and then they are divided into four categories: (1) the means-based method, (2) the Gaussian-mixture-model-based method, (3) the entropy-based method and (4) the feature-points-based method. The purpose of this tutorial is threefold: 1) to teach the principles of the classical histogram-based image segmentation methods to the interested readers; 2) to evaluate the advantages and disadvantages of these classical histogram-based image segmentation methods objectively; 3) to compare the performances of these classical histogram-based image segmentation methods with state-of-the-art deep learning based methods objectively.

♻ ☆ LVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents

Boyu Chen, Zhengrong Yue, Siran Chen, Zikang Wang, Yang Liu, Peng Li, Yali Wang

Existing Multimodal Large Language Models (MLLMs) encounter significant challenges in modeling the temporal context within long videos. Currently, mainstream Agent-based methods use external tools (e.g., search engine, memory banks, OCR, retrieval models) to assist a single MLLM in answering long video questions. Despite such tool-based support, a solitary MLLM still offers only a partial understanding of long videos, resulting in limited performance. In order to better address long video tasks, we introduce LVAgent, the first framework enabling multi-round dynamic collaboration of MLLM agents in long video understanding. Our methodology consists of four key steps: 1. Selection: We pre-select appropriate agents from the model library to form optimal agent teams based on different tasks. 2. Perception: We design an effective retrieval scheme for long videos, improving the coverage of critical temporal segments while maintaining computational efficiency. 3. Action: Agents answer long video-related questions and exchange reasons. 4. Reflection: We evaluate the performance of each agent in each round of discussion and optimize the agent team for dynamic collaboration. The agents iteratively refine their answers by multi-round dynamical collaboration of MLLM agents. LVAgent is the first agent system method that outperforms all closed-source models (including GPT-4o) and open-source models (including InternVL-2.5 and Qwen2-VL) in the long video understanding tasks. Our LVAgent achieves an accuracy of 80% on four mainstream long video understanding tasks. Notably, on the LongVideoBench dataset, LVAgent improves accuracy by up to 13.3% compared with SOTA.

♻ ☆ RG-Attn: Radian Glue Attention for Multi-modality Multi-agent Cooperative Perception

Lantao Li, Kang Yang, Wenqi Zhang, Xiaoxue Wang, Chen Sun

Cooperative perception offers an optimal solution to overcome the perception limitations of single-agent systems by leveraging Vehicle-to-Everything (V2X) communication for data sharing and fusion across multiple agents. However, most existing approaches focus on single-modality data exchange, limiting the potential of both homogeneous and heterogeneous fusion across agents. This overlooks the opportunity to utilize multi-modality data per agent, restricting the system's performance. In the automotive industry, manufacturers adopt diverse sensor configurations, resulting in heterogeneous combinations of sensor modalities across agents. To harness the potential of every possible data source for optimal performance, we design a robust LiDAR and camera cross-modality fusion module, Radian-Glue-Attention (RG-Attn), applicable to both intra-agent cross-modality fusion and inter-agent cross-modality fusion scenarios, owing to the convenient coordinate conversion by transformation matrix and the unified sampling/inversion mechanism. We also propose two different architectures, named Paint-To-Puzzle (PTP) and Co-Sketching-Co-Coloring (CoS-CoCo), for conducting cooperative perception. PTP aims for maximum precision performance and achieves smaller data packet size by limiting cross-agent fusion to a single instance, but requiring all participants to be equipped with LiDAR. In contrast, CoS-CoCo supports agents with any configuration-LiDAR-only, camera-only, or LiDAR-camera-both, presenting more generalization ability. Our approach achieves state-of-the-art (SOTA) performance on both real and simulated cooperative perception datasets. The code is now available at GitHub.

♻ ☆ PSF-4D: A Progressive Sampling Framework for View Consistent 4D Editing

Hasan Iqbal, Nazmul Karim, Umar Khalid, Azib Farooq, Zichun Zhong, Chen Chen, Jing Hua

Instruction-guided generative models, especially those using text-to-image (T2I) and text-to-video (T2V) diffusion frameworks, have advanced the field of content editing in recent years. To extend these capabilities to 4D scene, we introduce a progressive sampling framework for 4D editing (PSF-4D) that ensures temporal and multi-view consistency by intuitively controlling the noise initialization during forward diffusion. For temporal coherence, we design a correlated Gaussian noise structure that links frames over time, allowing each frame to depend meaningfully on prior frames. Additionally, to ensure spatial consistency across views, we implement a cross-view noise model, which uses shared and independent noise components to balance commonalities and distinct details among different views. To further enhance spatial coherence, PSF-4D incorporates view-consistent iterative refinement, embedding view-aware information into the denoising process to ensure aligned edits across frames and views. Our approach enables high-quality 4D editing without relying on external models, addressing key challenges in previous methods. Through extensive evaluation on multiple benchmarks and multiple editing aspects (e.g., style transfer, multi-attribute editing, object removal, local editing, etc.), we show the effectiveness of our proposed method. Experimental results demonstrate that our proposed method outperforms state-of-the-art 4D editing methods in diverse benchmarks.

comment: 9 pages, 7 figures

♻ ☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems

Ooha Lakkadi Reddy

This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.

comment: 106 pages (42 main text, 6 references, 58 appendices). 21 figures, 4 tables in main text; 106 figures, 8 tables total. Code: https://github.com/oohalakkadi/ivc2tyc. Undergraduate thesis at Duke Kunshan University. Accepted for presentation at the 52nd International Conference for Computer Applications & Quantitative Methods in Archaeology (CAA 2025), Athens, Greece

♻ ☆ Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lens

Zhangqi Jiang, Junkai Chen, Beier Zhu, Tingjin Luo, Yankun Shen, Xu Yang

Hallucinations in Large Vision-Language Models (LVLMs) significantly undermine their reliability, motivating researchers to explore the causes of hallucination. However, most studies primarily focus on the language aspect rather than the visual. In this paper, we address how LVLMs process visual information and whether this process causes hallucination. Firstly, we use the attention lens to identify the stages at which LVLMs handle visual data, discovering that the middle layers are crucial. Moreover, we find that these layers can be further divided into two stages: ''visual information enrichment'' and ''semantic refinement'' which respectively propagate visual data to object tokens and interpret it through text. By analyzing attention patterns during the visual information enrichment stage, we find that real tokens consistently receive higher attention weights than hallucinated ones, serving as a strong indicator of hallucination. Further examination of multi-head attention maps reveals that hallucination tokens often result from heads interacting with inconsistent objects. Based on these insights, we propose a simple inference-time method that adjusts visual attention by integrating information across various heads. Extensive experiments demonstrate that this approach effectively mitigates hallucinations in mainstream LVLMs without additional training costs. Code is available at https://github.com/ZhangqiJiang07/middle_layers_indicating_hallucinations.

♻ ☆ View-Invariant Pixelwise Anomaly Detection in Multi-object Scenes with Adaptive View Synthesis

Subin Varghese, Vedhus Hoskere

Visual anomaly detection in the built environment is a valuable tool for applications such as infrastructure assessment, construction monitoring, security surveillance, and urban planning. Anomaly detection approaches are typically unsupervised and work by detecting deviations from an expected state where no assumptions are made exact type of deviation. Unsupervised pixel-level anomaly detection methods have been developed to successfully recognize and segment anomalies; however, existing techniques are designed for industrial settings with a fixed camera position. In the built environment, images are periodically captured by a camera operated manually or mounted on aerial or ground vehicles. The camera pose between successive collections may vary widely voiding a fundamental assumption in existing anomaly detection approaches. To address this gap, we introduce the problem of Scene Anomaly Detection (Scene AD), where the goal is to detect anomalies from two sets of images: one set without anomalies and one set that may or may not contain anomalies. No labeled semantic segmentation data are provided for training. We propose a novel network, OmniAD, to tackle Scene AD by refining the reverse distillation anomaly detection method, leading to a 40\% improvement in pixel-level anomaly detection. Additionally, we introduce two new data augmentation strategies that leverage novel view synthesis and camera localization to enhance generalization. We evaluate our approach both qualitatively and quantitatively on a new dataset, ToyCity the first Scene AD dataset featuring multiple objects as well as on the established single object centric dataset, MAD. Our method demonstrates marked improvement over baseline approaches, paving the way for robust anomaly detection in scenes with real-world camera pose variations commonly observed in the built environment. https://drags99.github.io/OmniAD/

♻ ☆ Convolutional Neural Networks Can (Meta-)Learn the Same-Different Relation

Max Gupta, Sunayana Rane, R. Thomas McCoy, Thomas L. Griffiths

While convolutional neural networks (CNNs) have come to match and exceed human performance in many settings, the tasks these models optimize for are largely constrained to the level of individual objects, such as classification and captioning. Humans remain vastly superior to CNNs in visual tasks involving relations, including the ability to identify two objects as `same' or `different'. A number of studies have shown that while CNNs can be coaxed into learning the same-different relation in some settings, they tend to generalize poorly to other instances of this relation. In this work we show that the same CNN architectures that fail to generalize the same-different relation with conventional training are able to succeed when trained via meta-learning, which explicitly encourages abstraction and generalization across tasks.

♻ ☆ Enhancing Domain Adaptation through Prompt Gradient Alignment NeurIPS 2024

Hoang Phan, Lam Tran, Quyen Tran, Trung Le

Prior Unsupervised Domain Adaptation (UDA) methods often aim to train a domain-invariant feature extractor, which may hinder the model from learning sufficiently discriminative features. To tackle this, a line of works based on prompt learning leverages the power of large-scale pre-trained vision-language models to learn both domain-invariant and specific features through a set of domain-agnostic and domain-specific learnable prompts. Those studies typically enforce invariant constraints on representation, output, or prompt space to learn such prompts. In contrast, we cast UDA as a multiple-objective optimization problem in which each objective is represented by a domain loss. Under this new framework, we propose to align per-objective gradients to foster consensus between them. Additionally, to prevent potential overfitting when fine-tuning this deep learning architecture, we penalize the norm of these gradients. To achieve these goals, we devise a practical gradient update procedure that can work under both single-source and multi-source UDA. Empirically, our method consistently outperforms other vision-language model adaptation methods. The implementation is available at https://github.com/VietHoang1512/PGA.

comment: Accepted to NeurIPS 2024

♻ ☆ Disentangling Safe and Unsafe Corruptions via Anisotropy and Locality

Ramchandran Muthukumar, Ambar Pal, Jeremias Sulam, Rene Vidal

State-of-the-art machine learning systems are vulnerable to small perturbations to their input, where ``small'' is defined according to a threat model that assigns a positive threat to each perturbation. Most prior works define a task-agnostic, isotropic, and global threat, like the $\ell_p$ norm, where the magnitude of the perturbation fully determines the degree of the threat and neither the direction of the attack nor its position in space matter. However, common corruptions in computer vision, such as blur, compression, or occlusions, are not well captured by such threat models. This paper proposes a novel threat model called \texttt{Projected Displacement} (PD) to study robustness beyond existing isotropic and global threat models. The proposed threat model measures the threat of a perturbation via its alignment with \textit{unsafe directions}, defined as directions in the input space along which a perturbation of sufficient magnitude changes the ground truth class label. Unsafe directions are identified locally for each input based on observed training data. In this way, the PD threat model exhibits anisotropy and locality. Experiments on Imagenet-1k data indicate that, for any input, the set of perturbations with small PD threat includes \textit{safe} perturbations of large $\ell_p$ norm that preserve the true label, such as noise, blur and compression, while simultaneously excluding \textit{unsafe} perturbations that alter the true label. Unlike perceptual threat models based on embeddings of large-vision models, the PD threat model can be readily computed for arbitrary classification tasks without pre-training or finetuning. Further additional task annotation such as sensitivity to image regions or concept hierarchies can be easily integrated into the assessment of threat and thus the PD threat model presents practitioners with a flexible, task-driven threat specification.

comment: Published at IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025. Updated Acknowledgements

♻ ☆ NTSEBENCH: Cognitive Reasoning Benchmark for Vision Language Models

Pranshu Pandya, Vatsal Gupta, Agney S Talwarr, Tushar Kataria, Dan Roth, Vivek Gupta

Cognitive textual and visual reasoning tasks, including puzzles, series, and analogies, demand the ability to quickly reason, decipher, and evaluate patterns both textually and spatially. Due to extensive training on vast amounts of human-curated data, LLMs and VLMs excel in common-sense reasoning tasks, however still struggle with more complex reasoning that demands deeper cognitive understanding. We introduce NTSEBench, a new dataset designed to evaluate cognitive multi-modal reasoning and problem-solving skills of large models. The dataset contains 2728 multiple-choice questions, accompanied by a total of 4,642 images, categorized into 26 different types. These questions are drawn from the nationwide NTSE examination in India and feature a mix of visual and textual general aptitude challenges, designed to assess intelligence and critical thinking skills beyond mere rote learning. We establish baselines on the dataset using state-of-the-art LLMs and VLMs. To facilitate a comparison between open source and propriety models, we propose four distinct modeling strategies to handle different modalities -- text and images -- in the dataset instances.

comment: 28 pages, 3 figures, 12 tables

♻ ☆ Modeling and Analyzing the Influence of Non-Item Pages on Sequential Next-Item Prediction

Elisabeth Fischer, Albin Zehe, Andreas Hotho, Daniel Schlör

Analyzing sequences of interactions between users and items, sequential recommendation models can learn user intent and make predictions about the next item. Next to item interactions, most systems also have interactions with what we call non-item pages: these pages are not related to specific items but still can provide insights into the user's interests, as, for example, navigation pages. We therefore propose a general way to include these non-item pages in sequential recommendation models to enhance next-item prediction. First, we demonstrate the influence of non-item pages on following interactions using the hypotheses testing framework HypTrails and propose methods for representing non-item pages in sequential recommendation models. Subsequently, we adapt popular sequential recommender models to integrate non-item pages and investigate their performance with different item representation strategies as well as their ability to handle noisy data. To show the general capabilities of the models to integrate non-item pages, we create a synthetic dataset for a controlled setting and then evaluate the improvements from including non-item pages on two real-world datasets. Our results show that non-item pages are a valuable source of information, and incorporating them in sequential recommendation models increases the performance of next-item prediction across all analyzed model architectures.

comment: 40 pages, 19 figures; Accepted for ACM TORS Journal, Updated copyright information

♻ ☆ Knowledge-Aware Iterative Retrieval for Multi-Agent Systems

Seyoung Song

We introduce a novel large language model (LLM)-driven agent framework, which iteratively refines queries and filters contextual evidence by leveraging dynamically evolving knowledge. A defining feature of the system is its decoupling of external sources from an internal knowledge cache that is progressively updated to guide both query generation and evidence selection. This design mitigates bias-reinforcement loops and enables dynamic, trackable search exploration paths, thereby optimizing the trade-off between exploring diverse information and maintaining accuracy through autonomous agent decision-making. Our approach is evaluated on a broad range of open-domain question answering benchmarks, including multi-step tasks that mirror real-world scenarios where integrating information from multiple sources is critical, especially given the vulnerabilities of LLMs that lack explicit reasoning or planning capabilities. The results show that the proposed system not only outperforms single-step baselines regardless of task difficulty but also, compared to conventional iterative retrieval methods, demonstrates pronounced advantages in complex tasks through precise evidence-based reasoning and enhanced efficiency. The proposed system supports both competitive and collaborative sharing of updated context, enabling multi-agent extension. The benefits of multi-agent configurations become especially prominent as task difficulty increases. The number of convergence steps scales with task difficulty, suggesting cost-effective scalability.

♻ ☆ TOBUGraph: Knowledge Graph-Based Retrieval for Enhanced LLM Performance Beyond RAG

Savini Kashmira, Jayanaka L. Dantanarayana, Joshua Brodsky, Ashish Mahendra, Yiping Kang, Krisztian Flautner, Lingjia Tang, Jason Mars

Retrieval-Augmented Generation (RAG) is one of the leading and most widely used techniques for enhancing LLM retrieval capabilities, but it still faces significant limitations in commercial use cases. RAG primarily relies on the query-chunk text-to-text similarity in the embedding space for retrieval and can fail to capture deeper semantic relationships across chunks, is highly sensitive to chunking strategies, and is prone to hallucinations. To address these challenges, we propose TOBUGraph, a graph-based retrieval framework that first constructs the knowledge graph from unstructured data dynamically and automatically. Using LLMs, TOBUGraph extracts structured knowledge and diverse relationships among data, going beyond RAG's text-to-text similarity. Retrieval is achieved through graph traversal, leveraging the extracted relationships and structures to enhance retrieval accuracy, eliminating the need for chunking configurations while reducing hallucination. We demonstrate TOBUGraph's effectiveness in TOBU, a real-world application in production for personal memory organization and retrieval. Our evaluation using real user data demonstrates that TOBUGraph outperforms multiple RAG implementations in both precision and recall, significantly improving user experience through improved retrieval accuracy.

♻ ☆ FsPONER: Few-shot Prompt Optimization for Named Entity Recognition in Domain-specific Scenarios ECAI-2024

Yongjian Tang, Rakebul Hasan, Thomas Runkler

Large Language Models (LLMs) have provided a new pathway for Named Entity Recognition (NER) tasks. Compared with fine-tuning, LLM-powered prompting methods avoid the need for training, conserve substantial computational resources, and rely on minimal annotated data. Previous studies have achieved comparable performance to fully supervised BERT-based fine-tuning approaches on general NER benchmarks. However, none of the previous approaches has investigated the efficiency of LLM-based few-shot learning in domain-specific scenarios. To address this gap, we introduce FsPONER, a novel approach for optimizing few-shot prompts, and evaluate its performance on domain-specific NER datasets, with a focus on industrial manufacturing and maintenance, while using multiple LLMs -- GPT-4-32K, GPT-3.5-Turbo, LLaMA 2-chat, and Vicuna. FsPONER consists of three few-shot selection methods based on random sampling, TF-IDF vectors, and a combination of both. We compare these methods with a general-purpose GPT-NER method as the number of few-shot examples increases and evaluate their optimal NER performance against fine-tuned BERT and LLaMA 2-chat. In the considered real-world scenarios with data scarcity, FsPONER with TF-IDF surpasses fine-tuned models by approximately 10% in F1 score.

comment: accepted in the main track at the 27th European Conference on Artificial Intelligence (ECAI-2024)

♻ ☆ GME: Improving Universal Multimodal Retrieval by Multimodal LLMs CVPR 2025

Xin Zhang, Yanzhao Zhang, Wen Xie, Mingxin Li, Ziqi Dai, Dingkun Long, Pengjun Xie, Meishan Zhang, Wenjie Li, Min Zhang

Universal Multimodal Retrieval (UMR) aims to enable search across various modalities using a unified model, where queries and candidates can consist of pure text, images, or a combination of both. Previous work has attempted to adopt multimodal large language models (MLLMs) to realize UMR using only text data. However, our preliminary experiments demonstrate that more diverse multimodal training data can further unlock the potential of MLLMs. Despite its effectiveness, the existing multimodal training data is highly imbalanced in terms of modality, which motivates us to develop a training data synthesis pipeline and construct a large-scale, high-quality fused-modal training dataset. Based on the synthetic training data, we develop the General Multimodal Embedder (GME), an MLLM-based dense retriever designed for UMR. Furthermore, we construct a comprehensive UMR Benchmark (UMRB) to evaluate the effectiveness of our approach. Experimental results show that our method achieves state-of-the-art performance among existing UMR methods. Last, we provide in-depth analyses of model scaling and training strategies, and perform ablation studies on both the model and synthetic data.

comment: Accepted to CVPR 2025, models at https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-2B-Instruct

♻ ☆ UQABench: Evaluating User Embedding for Prompting LLMs in Personalized Question Answering

Langming Liu, Shilei Liu, Yujin Yuan, Yizhen Zhang, Bencheng Yan, Zhiyuan Zeng, Zihao Wang, Jiaqi Liu, Di Wang, Wenbo Su, Pengjie Wang, Jian Xu, Bo Zheng

Large language models (LLMs) achieve remarkable success in natural language processing (NLP). In practical scenarios like recommendations, as users increasingly seek personalized experiences, it becomes crucial to incorporate user interaction history into the context of LLMs to enhance personalization. However, from a practical utility perspective, user interactions' extensive length and noise present challenges when used directly as text prompts. A promising solution is to compress and distill interactions into compact embeddings, serving as soft prompts to assist LLMs in generating personalized responses. Although this approach brings efficiency, a critical concern emerges: Can user embeddings adequately capture valuable information and prompt LLMs? To address this concern, we propose \name, a benchmark designed to evaluate the effectiveness of user embeddings in prompting LLMs for personalization. We establish a fair and standardized evaluation process, encompassing pre-training, fine-tuning, and evaluation stages. To thoroughly evaluate user embeddings, we design three dimensions of tasks: sequence understanding, action prediction, and interest perception. These evaluation tasks cover the industry's demands in traditional recommendation tasks, such as improving prediction accuracy, and its aspirations for LLM-based methods, such as accurately understanding user interests and enhancing the user experience. We conduct extensive experiments on various state-of-the-art methods for modeling user embeddings. Additionally, we reveal the scaling laws of leveraging user embeddings to prompt LLMs. The benchmark is available online.

comment: 10 pages, 3 figures, 7 tables

♻ ☆ Robust Recommender System: A Survey and Future Directions

Kaike Zhang, Qi Cao, Fei Sun, Yunfan Wu, Shuchang Tao, Huawei Shen, Xueqi Cheng

With the rapid growth of information, recommender systems have become integral for providing personalized suggestions and overcoming information overload. However, their practical deployment often encounters ``dirty'' data, where noise or malicious information can lead to abnormal recommendations. Research on improving recommender systems' robustness against such dirty data has thus gained significant attention. This survey provides a comprehensive review of recent work on recommender systems' robustness. We first present a taxonomy to organize current techniques for withstanding malicious attacks and natural noise. We then explore state-of-the-art methods in each category, including fraudster detection, adversarial training, certifiable robust training for defending against malicious attacks, and regularization, purification, self-supervised learning for defending against malicious attacks. Additionally, we summarize evaluation metrics and commonly used datasets for assessing robustness. We discuss robustness across varying recommendation scenarios and its interplay with other properties like accuracy, interpretability, privacy, and fairness. Finally, we delve into open issues and future research directions in this emerging field. Our goal is to provide readers with a comprehensive understanding of robust recommender systems and to identify key pathways for future research and development. To facilitate ongoing exploration, we maintain a continuously updated GitHub repository with related research: https://github.com/Kaike-Zhang/Robust-Recommender-System.

♻ ☆ Exploring Multidimensional Checkworthiness: Designing AI-assisted Claim Prioritization for Human Fact-checkers SC

Houjiang Liu, Jacek Gwizdka, Matthew Lease

Given the massive volume of potentially false claims circulating online, claim prioritization is essential in allocating limited human resources available for fact-checking. In this study, we perceive claim prioritization as an information retrieval (IR) task: just as multidimensional IR relevance, with many factors influencing which search results a user deems relevant, checkworthiness is also multi-faceted, subjective, and even personal, with many factors influencing how fact-checkers triage and select which claims to check. Our study investigated both the multidimensional nature of checkworthiness and effective tool support to assist fact-checkers in claim prioritization. Methodologically, we pursued Research through Design combined with mixed-method evaluation. Specifically, we developed an AI-assisted claim prioritization prototype as a probe to explore how fact-checkers use multidimensional checkworthy factors to prioritize claims, simultaneously probing fact-checker needs and exploring the design space to meet those needs. With 16 professional fact-checkers participating in our study, we uncovered a hierarchical prioritization strategy fact-checkers implicitly use, revealing an underexplored aspect of their workflow, with actionable design recommendations for improving claim triage across multidimensional checkworthiness and tailoring this process with LLM integration.

comment: Accepted at CSCW 2025

♻ ☆ CoRanking: Collaborative Ranking with Small and Large Ranking Agents

Wenhan Liu, Xinyu Ma, Yutao Zhu, Lixin Su, Shuaiqiang Wang, Dawei Yin, Zhicheng Dou

Large Language Models (LLMs) have demonstrated superior listwise ranking performance. However, their superior performance often relies on large-scale parameters (\eg, GPT-4) and a repetitive sliding window process, which introduces significant efficiency challenges. In this paper, we propose \textbf{CoRanking}, a novel collaborative ranking framework that combines small and large ranking models for efficient and effective ranking. CoRanking first employs a small-size reranker to pre-rank all the candidate passages, bringing relevant ones to the top part of the list (\eg, top-20). Then, the LLM listwise reranker is applied to only rerank these top-ranked passages instead of the whole list, substantially enhancing overall ranking efficiency. Although more efficient, previous studies have revealed that the LLM listwise reranker have significant positional biases on the order of input passages. Directly feed the top-ranked passages from small reranker may result in the sub-optimal performance of LLM listwise reranker. To alleviate this problem, we introduce a passage order adjuster trained via reinforcement learning, which reorders the top passages from the small reranker to align with the LLM's preferences of passage order. Extensive experiments on three IR benchmarks demonstrate that CoRanking significantly improves efficiency (reducing ranking latency by about 70\%) while achieving even better effectiveness compared to using only the LLM listwise reranker.

♻ ☆ Low-Rank Thinning

Annabelle Michael Carrell, Albert Gong, Abhishek Shetty, Raaz Dwivedi, Lester Mackey

The goal in thinning is to summarize a dataset using a small set of representative points. Remarkably, sub-Gaussian thinning algorithms like Kernel Halving and Compress can match the quality of uniform subsampling while substantially reducing the number of summary points. However, existing guarantees cover only a restricted range of distributions and kernel-based quality measures and suffer from pessimistic dimension dependence. To address these deficiencies, we introduce a new low-rank analysis of sub-Gaussian thinning that applies to any distribution and any kernel, guaranteeing high-quality compression whenever the kernel or data matrix is approximately low-rank. To demonstrate the broad applicability of the techniques, we design practical sub-Gaussian thinning approaches that improve upon the best known guarantees for approximating attention in transformers, accelerating stochastic gradient training through reordering, and distinguishing distributions in near-linear time.

♻ ☆ Scalable Mechanistic Neural Networks for Differential Equations and Machine Learning ICLR 2025

Jiale Chen, Dingling Yao, Adeel Pervez, Dan Alistarh, Francesco Locatello

We propose Scalable Mechanistic Neural Network (S-MNN), an enhanced neural network framework designed for scientific machine learning applications involving long temporal sequences. By reformulating the original Mechanistic Neural Network (MNN) (Pervez et al., 2024), we reduce the computational time and space complexities from cubic and quadratic with respect to the sequence length, respectively, to linear. This significant improvement enables efficient modeling of long-term dynamics without sacrificing accuracy or interpretability. Extensive experiments demonstrate that S-MNN matches the original MNN in precision while substantially reducing computational resources. Consequently, S-MNN can drop-in replace the original MNN in applications, providing a practical and efficient tool for integrating mechanistic bottlenecks into neural network models of complex dynamical systems. Source code is available at https://github.com/IST-DASLab/ScalableMNN.

comment: Published as a conference paper at the Thirteenth International Conference on Learning Representations (ICLR 2025): https://openreview.net/forum?id=Oazgf8A24z

♻ ☆ Large-Scale Multi-omic Biosequence Transformers for Modeling Protein-Nucleic Acid Interactions

Sully F. Chen, Robert J. Steele, Glen M. Hocky, Beakal Lemeneh, Shivanand P. Lad, Eric K. Oermann

The transformer architecture has revolutionized bioinformatics and driven progress in the understanding and prediction of the properties of biomolecules. Almost all research on large-scale biosequence transformers has focused on one domain at a time (single-omic), usually DNA/RNA or proteins. These models have seen incredible success in downstream tasks in each domain, and have achieved particularly noteworthy breakthroughs in sequence modeling and structural modeling. However, these single-omic models are naturally incapable of efficiently modeling multi-omic tasks, one of the most biologically critical being protein-nucleic acid interactions. We present our work training the largest open-source multi-omic foundation model to date. We show that these multi-omic models (MOMs) can learn joint representations between various single-omic distributions that are emergently consistent with the Central Dogma of molecular biology despite only being trained on unlabeled biosequences. We further demonstrate that MOMs can be fine-tuned to achieve state-of-the-art results on protein-nucleic acid interaction tasks, namely predicting the change in Gibbs free energy ($\Delta G$) of the binding interaction between a given nucleic acid and protein. Remarkably, we show that multi-omic biosequence transformers emergently learn useful structural information without any \textit{a priori} structural training, allowing us to predict which protein residues are most involved in the protein-nucleic acid binding interaction. Lastly, we provide evidence that multi-omic biosequence models are in many cases superior to foundation models trained on single-omics distributions, both in performance-per-FLOP and absolute performance, suggesting a more generalized or foundational approach to building these models for biology.

comment: 39 pages, 5 figures

♻ ☆ Rehearsal-free Federated Domain-incremental Learning

Rui Sun, Haoran Duan, Jiahua Dong, Varun Ojha, Tejal Shah, Rajiv Ranjan

We introduce a rehearsal-free federated domain incremental learning framework, RefFiL, based on a global prompt-sharing paradigm to alleviate catastrophic forgetting challenges in federated domain-incremental learning, where unseen domains are continually learned. Typical methods for mitigating forgetting, such as the use of additional datasets and the retention of private data from earlier tasks, are not viable in federated learning (FL) due to devices' limited resources. Our method, RefFiL, addresses this by learning domain-invariant knowledge and incorporating various domain-specific prompts from the domains represented by different FL participants. A key feature of RefFiL is the generation of local fine-grained prompts by our domain adaptive prompt generator, which effectively learns from local domain knowledge while maintaining distinctive boundaries on a global scale. We also introduce a domain-specific prompt contrastive learning loss that differentiates between locally generated prompts and those from other domains, enhancing RefFiL's precision and effectiveness. Compared to existing methods, RefFiL significantly alleviates catastrophic forgetting without requiring extra memory space, making it ideal for privacy-sensitive and resource-constrained devices.

comment: Camera ready version. Accepted by the IEEE ICDCS, 2025

♻ ☆ Identifying Predictions That Influence the Future: Detecting Performative Concept Drift in Data Streams AAAI2025

Brandon Gower-Winter, Georg Krempl, Sergey Dragomiretskiy, Tineke Jelsma, Arno Siebes

Concept Drift has been extensively studied within the context of Stream Learning. However, it is often assumed that the deployed model's predictions play no role in the concept drift the system experiences. Closer inspection reveals that this is not always the case. Automated trading might be prone to self-fulfilling feedback loops. Likewise, malicious entities might adapt to evade detectors in the adversarial setting resulting in a self-negating feedback loop that requires the deployed models to constantly retrain. Such settings where a model may induce concept drift are called performative. In this work, we investigate this phenomenon. Our contributions are as follows: First, we define performative drift within a stream learning setting and distinguish it from other causes of drift. We introduce a novel type of drift detection task, aimed at identifying potential performative concept drift in data streams. We propose a first such performative drift detection approach, called CheckerBoard Performative Drift Detection (CB-PDD). We apply CB-PDD to both synthetic and semi-synthetic datasets that exhibit varying degrees of self-fulfilling feedback loops. Results are positive with CB-PDD showing high efficacy, low false detection rates, resilience to intrinsic drift, comparability to other drift detection techniques, and an ability to effectively detect performative drift in semi-synthetic datasets. Secondly, we highlight the role intrinsic (traditional) drift plays in obfuscating performative drift and discuss the implications of these findings as well as the limitations of CB-PDD.

comment: 21 pages, 17 figures. Extended version of paper with the same name accepted to AAAI2025 v2.0 updated the figures and text to more align with conference paper. Acknowledgements Section added

♻ ☆ A Survey on Unlearnable Data

Jiahao Li, Yiqiang Chen, Yunbing Xing, Yang Gu, Xiangyuan Lan

Unlearnable data (ULD) has emerged as an innovative defense technique to prevent machine learning models from learning meaningful patterns from specific data, thus protecting data privacy and security. By introducing perturbations to the training data, ULD degrades model performance, making it difficult for unauthorized models to extract useful representations. Despite the growing significance of ULD, existing surveys predominantly focus on related fields, such as adversarial attacks and machine unlearning, with little attention given to ULD as an independent area of study. This survey fills that gap by offering a comprehensive review of ULD, examining unlearnable data generation methods, public benchmarks, evaluation metrics, theoretical foundations and practical applications. We compare and contrast different ULD approaches, analyzing their strengths, limitations, and trade-offs related to unlearnability, imperceptibility, efficiency and robustness. Moreover, we discuss key challenges, such as balancing perturbation imperceptibility with model degradation and the computational complexity of ULD generation. Finally, we highlight promising future research directions to advance the effectiveness and applicability of ULD, underscoring its potential to become a crucial tool in the evolving landscape of data protection in machine learning.

comment: 31 pages, 3 figures, Code in https://github.com/LiJiahao-Alex/Awesome-UnLearnable-Data

♻ ☆ Optimal generalisation and learning transition in extensive-width shallow neural networks near interpolation

Jean Barbier, Francesco Camilli, Minh-Toan Nguyen, Mauro Pastore, Rudy Skerk

We consider a teacher-student model of supervised learning with a fully-trained two-layer neural network whose width $k$ and input dimension $d$ are large and proportional. We provide an effective theory for approximating the Bayes-optimal generalisation error of the network for any activation function in the regime of sample size $n$ scaling quadratically with the input dimension, i.e., around the interpolation threshold where the number of trainable parameters $kd+k$ and of data $n$ are comparable. Our analysis tackles generic weight distributions. We uncover a discontinuous phase transition separating a "universal" phase from a "specialisation" phase. In the first, the generalisation error is independent of the weight distribution and decays slowly with the sampling rate $n/d^2$, with the student learning only some non-linear combinations of the teacher weights. In the latter, the error is weight distribution-dependent and decays faster due to the alignment of the student towards the teacher network. We thus unveil the existence of a highly predictive solution near interpolation, which is however potentially hard to find by practical algorithms.

comment: v2: 9 pages + appendix, 10 figures, 3 tables; added discussion on Gaussian inner weights (Fig. 2, 5 + Appendix H); added discussion on algorithmic complexity of specialisation (Appendix I and figures therein)

♻ ☆ NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals

Jaden Fiotto-Kaufman, Alexander R. Loftus, Eric Todd, Jannik Brinkmann, Koyena Pal, Dmitrii Troitskii, Michael Ripa, Adam Belfki, Can Rager, Caden Juang, Aaron Mueller, Samuel Marks, Arnab Sen Sharma, Francesca Lucchetti, Nikhil Prakash, Carla Brodley, Arjun Guha, Jonathan Bell, Byron C. Wallace, David Bau

We introduce NNsight and NDIF, technologies that work in tandem to enable scientific study of the representations and computations learned by very large neural networks. NNsight is an open-source system that extends PyTorch to introduce deferred remote execution. The National Deep Inference Fabric (NDIF) is a scalable inference service that executes NNsight requests, allowing users to share GPU resources and pretrained models. These technologies are enabled by the Intervention Graph, an architecture developed to decouple experimental design from model runtime. Together, this framework provides transparent and efficient access to the internals of deep neural networks such as very large language models (LLMs) without imposing the cost or complexity of hosting customized models individually. We conduct a quantitative survey of the machine learning literature that reveals a growing gap in the study of the internals of large-scale AI. We demonstrate the design and use of our framework to address this gap by enabling a range of research methods on huge models. Finally, we conduct benchmarks to compare performance with previous approaches. Code, documentation, and tutorials are available at https://nnsight.net/.

comment: Code at https://nnsight.net

♻ ☆ Optimizing Posterior Samples for Bayesian Optimization via Rootfinding ICLR 2025

Taiwo A. Adebiyi, Bach Do, Ruda Zhang

Bayesian optimization devolves the global optimization of a costly objective function to the global optimization of a sequence of acquisition functions. This inner-loop optimization can be catastrophically difficult if it involves posterior sample paths, especially in higher dimensions. We introduce an efficient global optimization strategy for posterior samples based on global rootfinding. It provides gradient-based optimizers with two sets of judiciously selected starting points, designed to combine exploration and exploitation. The number of starting points can be kept small without sacrificing optimization quality. Remarkably, even with just one point from each set, the global optimum is discovered most of the time. The algorithm scales practically linearly to high dimensions, breaking the curse of dimensionality. For Gaussian process Thompson sampling (GP-TS), we demonstrate remarkable improvement in both inner- and outer-loop optimization, surprisingly outperforming alternatives like EI and GP-UCB in most cases. Our approach also improves the performance of other posterior sample-based acquisition functions, such as variants of entropy search. Furthermore, we propose a sample-average formulation of GP-TS, which has a parameter to explicitly control exploitation and can be computed at the cost of one posterior sample. Our implementation is available at https://github.com/UQUH/TSRoots .

comment: Published at the Thirteenth International Conference on Learning Representations (ICLR 2025)

♻ ☆ Mixture of Experts Made Personalized: Federated Prompt Learning for Vision-Language Models ICLR 2025

Jun Luo, Chen Chen, Shandong Wu

Federated prompt learning benefits federated learning with CLIP-like Vision-Language Model's (VLM's) robust representation learning ability through prompt learning. However, current federated prompt learning methods are habitually restricted to the traditional FL paradigm, where the participating clients are generally only allowed to download a single globally aggregated model from the server. While justifiable for training full-sized models under federated settings, in this work, we argue that this paradigm is ill-suited for lightweight prompts. By facilitating the clients to download multiple pre-aggregated prompts as fixed non-local experts, we propose Personalized Federated Mixture of Adaptive Prompts (pFedMoAP), a novel FL framework that personalizes the prompt learning process through the lens of Mixture of Experts (MoE). pFedMoAP implements a local attention-based gating network that learns to generate enhanced text features for better alignment with local image data, benefiting from both local and downloaded non-local adaptive prompt experts. Extensive experiments on 9 datasets under various federated settings demonstrate the efficacy of the proposed pFedMoAP algorithm. The code is available at https://github.com/ljaiverson/pFedMoAP.

comment: ICLR 2025

♻ ☆ FastLloyd: Federated, Accurate, Secure, and Tunable $k$-Means Clustering with Differential Privacy

Abdulrahman Diaa, Thomas Humphries, Florian Kerschbaum

We study the problem of privacy-preserving $k$-means clustering in the horizontally federated setting. Existing federated approaches using secure computation suffer from substantial overheads and do not offer output privacy. At the same time, differentially private (DP) $k$-means algorithms either assume a trusted central curator or significantly degrade utility by adding noise in the local DP model. Naively combining the secure and central DP solutions results in a protocol with impractical overhead. Instead, our work provides enhancements to both the DP and secure computation components, resulting in a design that is faster, more private, and more accurate than previous work. By utilizing the computational DP model, we design a lightweight, secure aggregation-based approach that achieves five orders of magnitude speed-up over state-of-the-art related work. Furthermore, we not only maintain the utility of the state-of-the-art in the central model of DP, but we improve the utility further by designing a new DP clustering mechanism.

♻ ☆ Optimal or Greedy Decision Trees? Revisiting their Objectives, Tuning, and Performance

Jacobus G. M. van der Linden, Daniël Vos, Mathijs M. de Weerdt, Sicco Verwer, Emir Demirović

Recently there has been a surge of interest in optimal decision tree (ODT) methods that globally optimize accuracy directly, in contrast to traditional approaches that locally optimize an impurity or information metric. However, the value of optimal methods is not well understood yet, as the literature provides conflicting results, with some demonstrating superior out-of-sample performance of ODTs over greedy approaches, while others show the opposite. Through a novel extensive experimental study, we provide new insights into the design and behavior of learning decision trees. In particular, we identify and analyze two relatively unexplored aspects of ODTs: the objective function used in training trees, and tuning techniques. Thus, we address these three questions: what objective to optimize in ODTs; how to tune ODTs; and how do optimal and greedy methods compare? Our experimental evaluation examines 11 objective functions, six tuning methods, and six claims from the literature on optimal and greedy methods on 180 real and synthetic data sets. Through our analysis, both conceptually and experimentally, we show the effect of (non-)concave objectives in greedy and optimal approaches; we highlight the importance of proper tuning of ODTs; support and refute several claims from the literature; provide clear recommendations for researchers and practitioners on the usage of greedy and optimal methods; and code for future comparisons.

♻ ☆ One Policy to Run Them All: an End-to-end Learning Approach to Multi-Embodiment Locomotion

Nico Bohlinger, Grzegorz Czechmanowski, Maciej Krupka, Piotr Kicki, Krzysztof Walas, Jan Peters, Davide Tateo

Deep Reinforcement Learning techniques are achieving state-of-the-art results in robust legged locomotion. While there exists a wide variety of legged platforms such as quadruped, humanoids, and hexapods, the field is still missing a single learning framework that can control all these different embodiments easily and effectively and possibly transfer, zero or few-shot, to unseen robot embodiments. We introduce URMA, the Unified Robot Morphology Architecture, to close this gap. Our framework brings the end-to-end Multi-Task Reinforcement Learning approach to the realm of legged robots, enabling the learned policy to control any type of robot morphology. The key idea of our method is to allow the network to learn an abstract locomotion controller that can be seamlessly shared between embodiments thanks to our morphology-agnostic encoders and decoders. This flexible architecture can be seen as a potential first step in building a foundation model for legged robot locomotion. Our experiments show that URMA can learn a locomotion policy on multiple embodiments that can be easily transferred to unseen robot platforms in simulation and the real world.

♻ ☆ Explainable Bayesian Optimization

Tanmay Chakraborty, Christian Wirth, Christin Seifert

Manual parameter tuning of cyber-physical systems is a common practice, but it is labor-intensive. Bayesian Optimization (BO) offers an automated alternative, yet its black-box nature reduces trust and limits human-BO collaborative system tuning. Experts struggle to interpret BO recommendations due to the lack of explanations. This paper addresses the post-hoc BO explainability problem for cyber-physical systems. We introduce TNTRules (Tune-No-Tune Rules), a novel algorithm that provides both global and local explanations for BO recommendations. TNTRules generates actionable rules and visual graphs, identifying optimal solution bounds and ranges, as well as potential alternative solutions. Unlike existing explainable AI (XAI) methods, TNTRules is tailored specifically for BO, by encoding uncertainty via a variance pruning technique and hierarchical agglomerative clustering. A multi-objective optimization approach allows maximizing explanation quality. We evaluate TNTRules using established XAI metrics (Correctness, Completeness, and Compactness) and compare it against adapted baseline methods. The results demonstrate that TNTRules generates high-fidelity, compact, and complete explanations, significantly outperforming three baselines on 5 multi-objective testing functions and 2 hyperparameter tuning problems.

♻ ☆ Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond

Liang Wen, Yunke Cai, Fenrui Xiao, Xin He, Qi An, Zhenyu Duan, Yimin Du, Junchen Liu, Lifu Tang, Xiaowei Lv, Haosheng Zou, Yongchao Deng, Shousheng Jia, Xiangzheng Zhang

This paper introduces Light-R1, an open-source suite for training long reasoning models using reproducible and cost-effective methodology. Given the proprietary nature of data used in the DeepSeek-R1 series, we develop an alternative approach leveraging exclusively public data and models. Our curriculum training progressively increases data difficulty, combined with multi-staged post-training. Our Light-R1-32B model, trained from Qwen2.5-32B-Instruct, outperforms DeepSeek-R1-Distill-Qwen-32B in math reasoning. Experimental results show that this curriculum approach becomes more effective when distinct, diverse datasets are available for different training stages: fine-tuning DeepSeek-R1-Distilled models (pre-tuned by DeepSeek team on proprietary data) with 3,000 challenging examples from our curriculum dataset yielded state-of-the-art 7B and 14B models, while the 32B model, Light-R1-32B-DS performed comparably to QwQ-32B and DeepSeek-R1. Furthermore, we extend our work by applying GRPO on long reasoning models. Our final Light-R1-14B-DS achieves SOTA performance among 14B models in math, with AIME24 \& 25 scores of 74.0 and 60.2 respectively, surpassing many 32B models and DeepSeek-R1-Distill-Llama-70B. Despite math-focused training, Light-R1-14B-DS demonstrates strong cross-domain generalization. Light-R1 represents a significant advancement in making sophisticated reasoning models more accessible and implementable in real-world applications. Our models, training data and code have been made available at https://github.com/Qihoo360/Light-R1.

comment: v3: minor modifications; v2: better writing & format for later submission; all release at https://github.com/Qihoo360/Light-R1

♻ ☆ Modeling and Analyzing the Influence of Non-Item Pages on Sequential Next-Item Prediction

Elisabeth Fischer, Albin Zehe, Andreas Hotho, Daniel Schlör

Analyzing sequences of interactions between users and items, sequential recommendation models can learn user intent and make predictions about the next item. Next to item interactions, most systems also have interactions with what we call non-item pages: these pages are not related to specific items but still can provide insights into the user's interests, as, for example, navigation pages. We therefore propose a general way to include these non-item pages in sequential recommendation models to enhance next-item prediction. First, we demonstrate the influence of non-item pages on following interactions using the hypotheses testing framework HypTrails and propose methods for representing non-item pages in sequential recommendation models. Subsequently, we adapt popular sequential recommender models to integrate non-item pages and investigate their performance with different item representation strategies as well as their ability to handle noisy data. To show the general capabilities of the models to integrate non-item pages, we create a synthetic dataset for a controlled setting and then evaluate the improvements from including non-item pages on two real-world datasets. Our results show that non-item pages are a valuable source of information, and incorporating them in sequential recommendation models increases the performance of next-item prediction across all analyzed model architectures.

comment: 40 pages, 19 figures; Accepted for ACM TORS Journal, Updated copyright information

♻ ☆ Hierarchical Procedural Framework for Low-latency Robot-Assisted Hand-Object Interaction

Mingqi Yuan, Huijiang Wang, Kai-Fung Chu, Fumiya Iida, Bo Li, Wenjun Zeng

Advances in robotics have been driving the development of human-robot interaction (HRI) technologies. However, accurately perceiving human actions and achieving adaptive control remains a challenge in facilitating seamless coordination between human and robotic movements. In this paper, we propose a hierarchical procedural framework to enable dynamic robot-assisted hand-object interaction. An open-loop hierarchy leverages the computer vision (CV)-based 3D reconstruction of the human hand, based on which motion primitives have been designed to translate hand motions into robotic actions. The low-level coordination hierarchy fine-tunes the robot's action by using the continuously updated 3D hand models. Experimental validation demonstrates the effectiveness of the hierarchical control architecture. The adaptive coordination between human and robot behavior has achieved a delay of $\leq 0.3$ seconds in the tele-interaction scenario. A case study of ring-wearing tasks indicates the potential application of this work in assistive technologies such as healthcare and manufacturing.

comment: 6 pages, 5 figures

♻ ☆ Semantic Learning for Molecular Communication in Internet of Bio-Nano Things

Hanlin Cai, Ozgur B. Akan

Molecular communication (MC) provides a foundational framework for information transmission in the Internet of Bio-Nano Things (IoBNT), where efficiency and reliability are crucial. However, the inherent limitations of molecular channels, such as low transmission rates, noise, and intersymbol interference (ISI), limit their ability to support complex data transmission. This paper proposes an end-to-end semantic learning framework designed to optimize task-oriented molecular communication, with a focus on biomedical diagnostic tasks under resource-constrained conditions. The proposed framework employs a deep encoder-decoder architecture to efficiently extract, quantize, and decode semantic features, prioritizing taskrelevant semantic information to enhance diagnostic classification performance. Additionally, a probabilistic channel network is introduced to approximate molecular propagation dynamics, enabling gradient-based optimization for end-to-end learning. Experimental results demonstrate that the proposed semantic framework improves diagnostic accuracy by at least 25% compared to conventional JPEG compression with LDPC coding methods under resource-constrained communication scenarios.

comment: This work has been accepted as an abstract paper for presentation at the 9th Workshop on Molecular Communications (MolCom), April 2025

♻ ☆ Patient-specific prediction of glioblastoma growth via reduced order modeling and neural networks

D. Cerrone, D. Riccobelli, S. Gazzoni, P. Vitullo, F. Ballarin, J. Falco, F. Acerbi, A. Manzoni, P. Zunino, P. Ciarletta

Glioblastoma is among the most aggressive brain tumors in adults, characterized by patient-specific invasion patterns driven by the underlying brain microstructure. In this work, we present a proof-of-concept for a mathematical model of GBL growth, enabling real-time prediction and patient-specific parameter identification from longitudinal neuroimaging data. The framework exploits a diffuse-interface mathematical model to describe the tumor evolution and a reduced-order modeling strategy, relying on proper orthogonal decomposition, trained on synthetic data derived from patient-specific brain anatomies reconstructed from magnetic resonance imaging and diffusion tensor imaging. A neural network surrogate learns the inverse mapping from tumor evolution to model parameters, achieving significant computational speed-up while preserving high accuracy. To ensure robustness and interpretability, we perform both global and local sensitivity analyses, identifying the key biophysical parameters governing tumor dynamics and assessing the stability of the inverse problem solution. These results establish a methodological foundation for future clinical deployment of patient-specific digital twins in neuro-oncology.

♻ ☆ TOBUGraph: Knowledge Graph-Based Retrieval for Enhanced LLM Performance Beyond RAG

Savini Kashmira, Jayanaka L. Dantanarayana, Joshua Brodsky, Ashish Mahendra, Yiping Kang, Krisztian Flautner, Lingjia Tang, Jason Mars

Retrieval-Augmented Generation (RAG) is one of the leading and most widely used techniques for enhancing LLM retrieval capabilities, but it still faces significant limitations in commercial use cases. RAG primarily relies on the query-chunk text-to-text similarity in the embedding space for retrieval and can fail to capture deeper semantic relationships across chunks, is highly sensitive to chunking strategies, and is prone to hallucinations. To address these challenges, we propose TOBUGraph, a graph-based retrieval framework that first constructs the knowledge graph from unstructured data dynamically and automatically. Using LLMs, TOBUGraph extracts structured knowledge and diverse relationships among data, going beyond RAG's text-to-text similarity. Retrieval is achieved through graph traversal, leveraging the extracted relationships and structures to enhance retrieval accuracy, eliminating the need for chunking configurations while reducing hallucination. We demonstrate TOBUGraph's effectiveness in TOBU, a real-world application in production for personal memory organization and retrieval. Our evaluation using real user data demonstrates that TOBUGraph outperforms multiple RAG implementations in both precision and recall, significantly improving user experience through improved retrieval accuracy.

♻ ☆ Innovative LSGTime Model for Crime Spatiotemporal Prediction Based on MindSpore Framework

Zhenkai Qin, BaoZhong Wei, Caifeng Gao

With the acceleration of urbanization, the spatiotemporal characteristics of criminal activities have become increasingly complex. Accurate prediction of crime distribution is crucial for optimizing the allocation of police resources and preventing crime. This paper proposes LGSTime, a crime spatiotemporal prediction model that integrates Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and the Multi-head Sparse Self-attention mechanism. LSTM and GRU capture long-term dependencies in crime time series, such as seasonality and periodicity, through their unique gating mechanisms. The Multi-head Sparse Self-attention mechanism, on the other hand, focuses on both temporal and spatial features of criminal events simultaneously through parallel processing and sparsification techniques, significantly improving computational efficiency and prediction accuracy. The integrated model leverages the strengths of each technique to better handle complex spatiotemporal data. Experimental findings demonstrate that the model attains optimal performance across four real - world crime datasets. In comparison to the CNN model, it exhibits performance enhancements of 2.8\%, 1.9\%, and 1.4\% in the Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) metrics respectively. These results offer a valuable reference for tackling the challenges in crime prediction.

♻ ☆ Improving Vector-Quantized Image Modeling with Latent Consistency-Matching Diffusion

Bac Nguyen, Chieh-Hsin Lai, Yuhta Takida, Naoki Murata, Toshimitsu Uesaka, Stefano Ermon, Yuki Mitsufuji

By embedding discrete representations into a continuous latent space, we can leverage continuous-space latent diffusion models to handle generative modeling of discrete data. However, despite their initial success, most latent diffusion methods rely on fixed pretrained embeddings, limiting the benefits of joint training with the diffusion model. While jointly learning the embedding (via reconstruction loss) and the latent diffusion model (via score matching loss) could enhance performance, end-to-end training risks embedding collapse, degrading generation quality. To mitigate this issue, we introduce VQ-LCMD, a continuous-space latent diffusion framework within the embedding space that stabilizes training. VQ-LCMD uses a novel training objective combining the joint embedding-diffusion variational lower bound with a consistency-matching (CM) loss, alongside a shifted cosine noise schedule and random dropping strategy. Experiments on several benchmarks show that the proposed VQ-LCMD yields superior results on FFHQ, LSUN Churches, and LSUN Bedrooms compared to discrete-state latent diffusion models. In particular, VQ-LCMD achieves an FID of 6.81 for class-conditional image generation on ImageNet with 50 steps.

♻ ☆ Statistically Testing Training Data for Unwanted Error Patterns using Rule-Oriented Regression

Stefan Rass, Martin Dallinger

Artificial intelligence models trained from data can only be as good as the underlying data is. Biases in training data propagating through to the output of a machine learning model are a well-documented and well-understood phenomenon, but the machinery to prevent these undesired effects is much less developed. Efforts to ensure data is clean during collection, such as using bias-aware sampling, are most effective when the entity controlling data collection also trains the AI. In cases where the data is already available, how do we find out if the data was already manipulated, i.e., ``poisoned'', so that an undesired behavior would be trained into a machine learning model? This is a challenge fundamentally different to (just) improving approximation accuracy or efficiency, and we provide a method to test training data for flaws, to establish a trustworthy ground-truth for a subsequent training of machine learning models (of any kind). Unlike the well-studied problem of approximating data using fuzzy rules that are generated from the data, our method hinges on a prior definition of rules to happen before seeing the data to be tested. Therefore, the proposed method can also discover hidden error patterns, which may also have substantial influence. Our approach extends the abilities of conventional statistical testing by letting the ``test-condition'' be any Boolean condition to describe a pattern in the data, whose presence we wish to determine. The method puts fuzzy inference into a regression model, to get the best of the two: explainability from fuzzy logic with statistical properties and diagnostics from the regression, and finally also being applicable to ``small data'', hence not requiring large datasets as deep learning methods do. We provide an open source implementation for demonstration and experiments.

♻ ☆ MSCMNet: Multi-scale Semantic Correlation Mining for Visible-Infrared Person Re-Identification

Xuecheng Hua, Ke Cheng, Hu Lu, Juanjuan Tu, Yuanquan Wang, Shitong Wang

The main challenge in the Visible-Infrared Person Re-Identification (VI-ReID) task lies in how to extract discriminative features from different modalities for matching purposes. While the existing well works primarily focus on minimizing the modal discrepancies, the modality information can not thoroughly be leveraged. To solve this problem, a Multi-scale Semantic Correlation Mining network (MSCMNet) is proposed to comprehensively exploit semantic features at multiple scales and simultaneously reduce modality information loss as small as possible in feature extraction. The proposed network contains three novel components. Firstly, after taking into account the effective utilization of modality information, the Multi-scale Information Correlation Mining Block (MIMB) is designed to explore semantic correlations across multiple scales. Secondly, in order to enrich the semantic information that MIMB can utilize, a quadruple-stream feature extractor (QFE) with non-shared parameters is specifically designed to extract information from different dimensions of the dataset. Finally, the Quadruple Center Triplet Loss (QCT) is further proposed to address the information discrepancy in the comprehensive features. Extensive experiments on the SYSU-MM01, RegDB, and LLCM datasets demonstrate that the proposed MSCMNet achieves the greatest accuracy.

♻ ☆ Class-Dependent Perturbation Effects in Evaluating Time Series Attributions

Gregor Baer, Isel Grau, Chao Zhang, Pieter Van Gorp

As machine learning models become increasingly prevalent in time series applications, Explainable Artificial Intelligence (XAI) methods are essential for understanding their predictions. Within XAI, feature attribution methods aim to identify which input features contribute the most to a model's prediction, with their evaluation typically relying on perturbation-based metrics. Through systematic empirical analysis across multiple datasets, model architectures, and perturbation strategies, we reveal previously overlooked class-dependent effects in these metrics: they show varying effectiveness across classes, achieving strong results for some while remaining less sensitive to others. In particular, we find that the most effective perturbation strategies often demonstrate the most pronounced class differences. Our analysis suggests that these effects arise from the learned biases of classifiers, indicating that perturbation-based evaluation may reflect specific model behaviors rather than intrinsic attribution quality. We propose an evaluation framework with a class-aware penalty term to help assess and account for these effects in evaluating feature attributions, offering particular value for class-imbalanced datasets. Although our analysis focuses on time series classification, these class-dependent effects likely extend to other structured data domains where perturbation-based evaluation is common.

comment: Accepted at The World Conference on eXplainable Artificial Intelligence (XAI-2025)

♻ ☆ Optimization Insights into Deep Diagonal Linear Networks

Hippolyte Labarrière, Cesare Molinari, Lorenzo Rosasco, Silvia Villa, Cristian Vega

Overparameterized models trained with (stochastic) gradient descent are ubiquitous in modern machine learning. These large models achieve unprecedented performance on test data, but their theoretical understanding is still limited. In this paper, we take a step towards filling this gap by adopting an optimization perspective. More precisely, we study the implicit regularization properties of the gradient flow "algorithm" for estimating the parameters of a deep diagonal neural network. Our main contribution is showing that this gradient flow induces a mirror flow dynamic on the model, meaning that it is biased towards a specific solution of the problem depending on the initialization of the network. Along the way, we prove several properties of the trajectory.

♻ ☆ A stochastic gradient descent algorithm with random search directions

Eméric Gbaguidi

Stochastic coordinate descent algorithms are efficient methods in which each iterate is obtained by fixing most coordinates at their values from the current iteration, and approximately minimizing the objective with respect to the remaining coordinates. However, this approach is usually restricted to canonical basis vectors of $\mathbb{R}^d$. In this paper, we develop a new class of stochastic gradient descent algorithms with random search directions which uses the directional derivative of the gradient estimate following more general random vectors. We establish the almost sure convergence of these algorithms with decreasing step. We further investigate their central limit theorem and pay particular attention to analyze the impact of the search distributions on the asymptotic covariance matrix. We also provide non-asymptotic $\mathbb{L}^p$ rates of convergence.

♻ ☆ Calibration Strategies for Robust Causal Estimation: Theoretical and Empirical Insights on Propensity Score Based Estimators

Jan Rabenseifner, Sven Klaassen, Jannis Kueck, Philipp Bach

The partitioning of data for estimation and calibration critically impacts the performance of propensity score based estimators like inverse probability weighting (IPW) and double/debiased machine learning (DML) frameworks. We extend recent advances in calibration techniques for propensity score estimation, improving the robustness of propensity scores in challenging settings such as limited overlap, small sample sizes, or unbalanced data. Our contributions are twofold: First, we provide a theoretical analysis of the properties of calibrated estimators in the context of DML. To this end, we refine existing calibration frameworks for propensity score models, with a particular emphasis on the role of sample-splitting schemes in ensuring valid causal inference. Second, through extensive simulations, we show that calibration reduces variance of inverse-based propensity score estimators while also mitigating bias in IPW, even in small-sample regimes. Notably, calibration improves stability for flexible learners (e.g., gradient boosting) while preserving the doubly robust properties of DML. A key insight is that, even when methods perform well without calibration, incorporating a calibration step does not degrade performance, provided that an appropriate sample-splitting approach is chosen.

♻ ☆ Towards Robust and Parameter-Efficient Knowledge Unlearning for LLMs ICLR 2025

Sungmin Cha, Sungjun Cho, Dasol Hwang, Moontae Lee

Large Language Models (LLMs) have demonstrated strong reasoning and memorization capabilities via pretraining on massive textual corpora. However, this poses risk of privacy and copyright violations, highlighting the need for efficient machine unlearning methods that remove sensitive data without retraining from scratch. While Gradient Ascent (GA) is commonly used to unlearn by reducing the likelihood of generating unwanted content, it leads to unstable optimization and catastrophic forgetting of retrained knowledge. We find that combining GA with low-rank adaptation results in poor trade-offs between computational cost and generative performance. To address these challenges, we propose Low-rank Knowledge Unlearning (LoKU), a novel framework that enables robust and efficient unlearning for LLMs. First, we introduce Inverted Hinge Loss, which suppresses unwanted tokens while maintaining fluency by boosting the probability of the next most likely token. Second, we develop a data-adaptive initialization for LoRA adapters via low-rank approximation weighted with relative Fisher information, thereby focusing updates on parameters critical for removing targeted knowledge. Experiments on the Training Data Extraction Challenge dataset using GPT-Neo models as well as on the TOFU benchmark with Phi-1.5B and Llama2-7B models demonstrate that our approach effectively removes sensitive information while maintaining reasoning and generative capabilities with minimal impact. Our implementation can be found in https://github.com/csm9493/efficient-llm-unlearning.

comment: ICLR 2025 camera-ready version

♻ ☆ Reasoning-SQL: Reinforcement Learning with SQL Tailored Partial Rewards for Reasoning-Enhanced Text-to-SQL

Mohammadreza Pourreza, Shayan Talaei, Ruoxi Sun, Xingchen Wan, Hailong Li, Azalia Mirhoseini, Amin Saberi, Sercan "O. Arik

Text-to-SQL is a challenging task involving multiple reasoning-intensive subtasks, including natural language understanding, database schema comprehension, and precise SQL query formulation. Existing approaches often rely on handcrafted reasoning paths with inductive biases that can limit their overall effectiveness. Motivated by the recent success of reasoning-enhanced models such as DeepSeek R1 and OpenAI o1, which effectively leverage reward-driven self-exploration to enhance reasoning capabilities and generalization, we propose a novel set of partial rewards tailored specifically for the Text-to-SQL task. Our reward set includes schema-linking, AI feedback, n-gram similarity, and syntax check, explicitly designed to address the reward sparsity issue prevalent in reinforcement learning (RL). Leveraging group relative policy optimization (GRPO), our approach explicitly encourages large language models (LLMs) to develop intrinsic reasoning skills necessary for accurate SQL query generation. With models of different sizes, we demonstrate that RL-only training with our proposed rewards consistently achieves higher accuracy and superior generalization compared to supervised fine-tuning (SFT). Remarkably, our RL-trained 14B-parameter model significantly outperforms larger proprietary models, e.g. o3-mini by 4% and Gemini-1.5-Pro-002 by 3% on the BIRD benchmark. These highlight the efficacy of our proposed RL-training framework with partial rewards for enhancing both accuracy and reasoning capabilities in Text-to-SQL tasks.

comment: Mohammadreza Pourreza and Shayan Talaei contributed equally to this work

♻ ☆ Generative Data Assimilation of Sparse Weather Station Observations at Kilometer Scales

Peter Manshausen, Yair Cohen, Peter Harrington, Jaideep Pathak, Mike Pritchard, Piyush Garg, Morteza Mardani, Karthik Kashinath, Simon Byrne, Noah Brenowitz

Data assimilation of observational data into full atmospheric states is essential for weather forecast model initialization. Recently, methods for deep generative data assimilation have been proposed which allow for using new input data without retraining the model. They could also dramatically accelerate the costly data assimilation process used in operational regional weather models. Here, in a central US testbed, we demonstrate the viability of score-based data assimilation in the context of realistically complex km-scale weather. We train an unconditional diffusion model to generate snapshots of a state-of-the-art km-scale analysis product, the High Resolution Rapid Refresh. Then, using score-based data assimilation to incorporate sparse weather station data, the model produces maps of precipitation and surface winds. The generated fields display physically plausible structures, such as gust fronts, and sensitivity tests confirm learnt physics through multivariate relationships. Preliminary skill analysis shows the approach already outperforms a naive baseline of the High-Resolution Rapid Refresh system itself. By incorporating observations from 40 weather stations, 10% lower RMSEs on left-out stations are attained. Despite some lingering imperfections such as insufficiently disperse ensemble DA estimates, we find the results overall an encouraging proof of concept, and the first at km-scale. It is a ripe time to explore extensions that combine increasingly ambitious regional state generators with an increasing set of in situ, ground-based, and satellite remote sensing data streams.

comment: 22 pages, 9 figures

♻ ☆ SVInvNet: A Densely Connected Encoder-Decoder Architecture for Seismic Velocity Inversion

Mojtaba Najafi Khatounabad, Hacer Yalim Keles, Selma Kadioglu

This study presents a deep learning-based approach to seismic velocity inversion problem, focusing on both noisy and noiseless training datasets of varying sizes. Our Seismic Velocity Inversion Network (SVInvNet) introduces a novel architecture that contains a multi-connection encoder-decoder structure enhanced with dense blocks. This design is specifically tuned to effectively process time series data, which is essential for addressing the challenges of non-linear seismic velocity inversion. For training and testing, we created diverse seismic velocity models, including multi-layered, faulty, and salt dome categories. We also investigated how different kinds of ambient noise, both coherent and stochastic, and the size of the training dataset affect learning outcomes. SVInvNet is trained on datasets ranging from 750 to 6,000 samples and is tested using a large benchmark dataset of 12,000 samples. Despite its fewer parameters compared to the baseline model, SVInvNet achieves superior performance with this dataset. The performance of SVInvNet was further evaluated using the OpenFWI dataset and Marmousi-derived velocity models. The comparative analysis clearly reveals the effectiveness of the proposed model.

comment: This is the preprint of the accepted manuscript to appear in IEEE Transactions on Geoscience and Remote Sensing

♻ ☆ Self-Supervised Pretraining for Aerial Road Extraction

Rupert Polley, Sai Vignesh Abishek Deenadayalan, J. Marius Zöllner

Deep neural networks for aerial image segmentation require large amounts of labeled data, but high-quality aerial datasets with precise annotations are scarce and costly to produce. To address this limitation, we propose a self-supervised pretraining method that improves segmentation performance while reducing reliance on labeled data. Our approach uses inpainting-based pretraining, where the model learns to reconstruct missing regions in aerial images, capturing their inherent structure before being fine-tuned for road extraction. This method improves generalization, enhances robustness to domain shifts, and is invariant to model architecture and dataset choice. Experiments show that our pretraining significantly boosts segmentation accuracy, especially in low-data regimes, making it a scalable solution for aerial image analysis.

comment: Accepted at 36th IEEE Intelligent Vehicles Symposium (IV) 2025 Joint Workshop on Safety, Metrics and Benchmarks for Autonomous Driving

♻ ☆ Sharp Rates in Dependent Learning Theory: Avoiding Sample Size Deflation for the Square Loss

Ingvar Ziemann, Stephen Tu, George J. Pappas, Nikolai Matni

In this work, we study statistical learning with dependent ($\beta$-mixing) data and square loss in a hypothesis class $\mathscr{F}\subset L_{\Psi_p}$ where $\Psi_p$ is the norm $\|f\|_{\Psi_p} \triangleq \sup_{m\geq 1} m^{-1/p} \|f\|_{L^m} $ for some $p\in [2,\infty]$. Our inquiry is motivated by the search for a sharp noise interaction term, or variance proxy, in learning with dependent data. Absent any realizability assumption, typical non-asymptotic results exhibit variance proxies that are deflated multiplicatively by the mixing time of the underlying covariates process. We show that whenever the topologies of $L^2$ and $\Psi_p$ are comparable on our hypothesis class $\mathscr{F}$ -- that is, $\mathscr{F}$ is a weakly sub-Gaussian class: $\|f\|_{\Psi_p} \lesssim \|f\|_{L^2}^\eta$ for some $\eta\in (0,1]$ -- the empirical risk minimizer achieves a rate that only depends on the complexity of the class and second order statistics in its leading term. Our result holds whether the problem is realizable or not and we refer to this as a \emph{near mixing-free rate}, since direct dependence on mixing is relegated to an additive higher order term. We arrive at our result by combining the above notion of a weakly sub-Gaussian class with mixed tail generic chaining. This combination allows us to compute sharp, instance-optimal rates for a wide range of problems. Examples that satisfy our framework include sub-Gaussian linear regression, more general smoothly parameterized function classes, finite hypothesis classes, and bounded smoothness classes.

♻ ☆ DG-TTA: Out-of-domain Medical Image Segmentation through Augmentation and Descriptor-driven Domain Generalization and Test-Time Adaptation

Christian Weihsbach, Christian N. Kruse, Alexander Bigalke, Mattias P. Heinrich

Purpose: Applying pre-trained medical deep learning segmentation models on out-of-domain images often yields predictions of insufficient quality. In this study, we propose to use a powerful generalizing descriptor along with augmentation to enable domain-generalized pre-training and test-time adaptation, achieving high-quality segmentation in unseen domains. Materials and Methods: In this retrospective study five different publicly available datasets (2012 to 2022) including 3D CT and MRI images are used to evaluate segmentation performance in out-of-domain scenarios. The settings include abdominal, spine, and cardiac imaging. The data is randomly split into training and test samples. Domain-generalized pre-training on source data is used to obtain the best initial performance in the target domain. We introduce the combination of the generalizing SSC descriptor and GIN intensity augmentation for optimal generalization. Segmentation results are subsequently optimized at test time, where we propose to adapt the pre-trained models for every unseen scan with a consistency scheme using the same augmentation-descriptor combination. The segmentation is evaluated using Dice similarity and Hausdorff distance and the significance of improvements is tested with the Wilcoxon signed-rank test. Results: The proposed generalized pre-training and subsequent test-time adaptation improves model performance significantly in CT to MRI cross-domain prediction for abdominal (+46.2% and +28.2% Dice), spine (+72.9%), and cardiac (+14.2% and +55.7% Dice) scenarios (p<0.001). Conclusion: Our method enables optimal, independent usage of medical image source and target data and bridges domain gaps successfully with a compact and efficient methodology. Open-source code available at: https://github.com/multimodallearning/DG-TTA

♻ ☆ Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation

Han Xue, Jieji Ren, Wendi Chen, Gu Zhang, Yuan Fang, Guoying Gu, Huazhe Xu, Cewu Lu

Humans can accomplish complex contact-rich tasks using vision and touch, with highly reactive capabilities such as quick adjustments to environmental changes and adaptive control of contact forces; however, this remains challenging for robots. Existing visual imitation learning (IL) approaches rely on action chunking to model complex behaviors, which lacks the ability to respond instantly to real-time tactile feedback during the chunk execution. Furthermore, most teleoperation systems struggle to provide fine-grained tactile / force feedback, which limits the range of tasks that can be performed. To address these challenges, we introduce TactAR, a low-cost teleoperation system that provides real-time tactile feedback through Augmented Reality (AR), along with Reactive Diffusion Policy (RDP), a novel slow-fast visual-tactile imitation learning algorithm for learning contact-rich manipulation skills. RDP employs a two-level hierarchy: (1) a slow latent diffusion policy for predicting high-level action chunks in latent space at low frequency, (2) a fast asymmetric tokenizer for closed-loop tactile feedback control at high frequency. This design enables both complex trajectory modeling and quick reactive behavior within a unified framework. Through extensive evaluation across three challenging contact-rich tasks, RDP significantly improves performance compared to state-of-the-art visual IL baselines through rapid response to tactile / force feedback. Furthermore, experiments show that RDP is applicable across different tactile / force sensors. Code and videos are available on https://reactive-diffusion-policy.github.io.

♻ ☆ Nonhuman Primate Brain Tissue Segmentation Using a Transfer Learning Approach

Zhen Lin, Hongyu Yuan, Richard Barcus, Qing Lyu, Sucheta Chakravarty, Megan E. Lipford, Carol A. Shively, Suzanne Craft, Mohammad Kawas, Jeongchul Kim, Christopher T. Whitlow

Non-human primates (NHPs) serve as critical models for understanding human brain function and neurological disorders due to their close evolutionary relationship with humans. Accurate brain tissue segmentation in NHPs is critical for understanding neurological disorders, but challenging due to the scarcity of annotated NHP brain MRI datasets, the small size of the NHP brain, the limited resolution of available imaging data and the anatomical differences between human and NHP brains. To address these challenges, we propose a novel approach utilizing STU-Net with transfer learning to leverage knowledge transferred from human brain MRI data to enhance segmentation accuracy in the NHP brain MRI, particularly when training data is limited. The combination of STU-Net and transfer learning effectively delineates complex tissue boundaries and captures fine anatomical details specific to NHP brains. Notably, our method demonstrated improvement in segmenting small subcortical structures such as putamen and thalamus that are challenging to resolve with limited spatial resolution and tissue contrast, and achieved DSC of over 0.88, IoU over 0.8 and HD95 under 7. This study introduces a robust method for multi-class brain tissue segmentation in NHPs, potentially accelerating research in evolutionary neuroscience and preclinical studies of neurological disorders relevant to human health.

♻ ☆ Exact full-RSB SAT/UNSAT transition in infinitely wide two-layer neural networks

Brandon L. Annesi, Enrico M. Malatesta, Francesco Zamponi

We analyze the problem of storing random pattern-label associations using two classes of continuous non-convex weights models, namely the perceptron with negative margin and an infinite-width two-layer neural network with non-overlapping receptive fields and generic activation function. Using a full-RSB ansatz we compute the exact value of the SAT/UNSAT transition. Furthermore, in the case of the negative perceptron we show that the overlap distribution of typical states displays an overlap gap (a disconnected support) in certain regions of the phase diagram defined by the value of the margin and the density of patterns to be stored. This implies that some recent theorems that ensure convergence of Approximate Message Passing (AMP) based algorithms to capacity are not applicable. Finally, we show that Gradient Descent is not able to reach the maximal capacity, irrespectively of the presence of an overlap gap for typical states. This finding, similarly to what occurs in binary weight models, suggests that gradient-based algorithms are biased towards highly atypical states, whose inaccessibility determines the algorithmic threshold.

comment: 39 pages, 12 figures

♻ ☆ Illuminating the Diversity-Fitness Trade-Off in Black-Box Optimization

Maria Laura Santoni, Elena Raponi, Aneta Neumann, Frank Neumann, Mike Preuss, Carola Doerr

In real-world applications, users often favor structurally diverse design choices over one high-quality solution. It is hence important to consider more solutions that decision makers can compare and further explore based on additional criteria. Alongside the existing approaches of evolutionary diversity optimization, quality diversity, and multimodal optimization, this paper presents a fresh perspective on this challenge by considering the problem of identifying a fixed number of solutions with a pairwise distance above a specified threshold while maximizing their average quality. We obtain first insight into these objectives by performing a subset selection on the search trajectories of different well-established search heuristics, whether they have been specifically designed with diversity in mind or not. We emphasize that the main goal of our work is not to present a new algorithm but to understand the capability of off-the-shelf algorithms to quantify the trade-off between the minimum pairwise distance within batches of solutions and their average quality. We also analyze how this trade-off depends on the properties of the underlying optimization problem. A possibly surprising outcome of our empirical study is the observation that naive uniform random sampling establishes a very strong baseline for our problem, hardly ever outperformed by the search trajectories of the considered heuristics. We interpret these results as a motivation to develop algorithms tailored to produce diverse solutions of high average quality.

♻ ☆ Machine Unlearning Fails to Remove Data Poisoning Attacks ICLR 2025

Martin Pawelczyk, Jimmy Z. Di, Yiwei Lu, Ayush Sekhari, Gautam Kamath, Seth Neel

We revisit the efficacy of several practical methods for approximate machine unlearning developed for large-scale deep learning. In addition to complying with data deletion requests, one often-cited potential application for unlearning methods is to remove the effects of poisoned data. We experimentally demonstrate that, while existing unlearning methods have been demonstrated to be effective in a number of settings, they fail to remove the effects of data poisoning across a variety of types of poisoning attacks (indiscriminate, targeted, and a newly-introduced Gaussian poisoning attack) and models (image classifiers and LLMs); even when granted a relatively large compute budget. In order to precisely characterize unlearning efficacy, we introduce new evaluation metrics for unlearning based on data poisoning. Our results suggest that a broader perspective, including a wider variety of evaluations, are required to avoid a false sense of confidence in machine unlearning procedures for deep learning without provable guarantees. Moreover, while unlearning methods show some signs of being useful to efficiently remove poisoned data without having to retrain, our work suggests that these methods are not yet ``ready for prime time,'' and currently provide limited benefit over retraining.

comment: Published at ICLR 2025

♻ ☆ Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning

Gabriele Dominici, Pietro Barbiero, Mateo Espinosa Zarlenga, Alberto Termine, Martin Gjoreski, Giuseppe Marra, Marc Langheinrich

Causal opacity denotes the difficulty in understanding the "hidden" causal structure underlying the decisions of deep neural network (DNN) models. This leads to the inability to rely on and verify state-of-the-art DNN-based systems, especially in high-stakes scenarios. For this reason, circumventing causal opacity in DNNs represents a key open challenge at the intersection of deep learning, interpretability, and causality. This work addresses this gap by introducing Causal Concept Graph Models (Causal CGMs), a class of interpretable models whose decision-making process is causally transparent by design. Our experiments show that Causal CGMs can: (i) match the generalisation performance of causally opaque models, (ii) enable human-in-the-loop corrections to mispredicted intermediate reasoning steps, boosting not just downstream accuracy after corrections but also the reliability of the explanations provided for specific instances, and (iii) support the analysis of interventional and counterfactual scenarios, thereby improving the model's causal interpretability and supporting the effective verification of its reliability and fairness.

♻ ☆ Leveraging Joint Predictive Embedding and Bayesian Inference in Graph Self Supervised Learning

Srinitish Srinivasan, Omkumar CU

Graph representation learning has emerged as a cornerstone for tasks like node classification and link prediction, yet prevailing self-supervised learning (SSL) methods face challenges such as computational inefficiency, reliance on contrastive objectives, and representation collapse. Existing approaches often depend on feature reconstruction, negative sampling, or complex decoders, which introduce training overhead and hinder generalization. Further, current techniques which address such limitations fail to account for the contribution of node embeddings to a certain prediction in the absence of labeled nodes. To address these limitations, we propose a novel joint embedding predictive framework for graph SSL that eliminates contrastive objectives and negative sampling while preserving semantic and structural information. Additionally, we introduce a semantic-aware objective term that incorporates pseudo-labels derived from Gaussian Mixture Models (GMMs), enhancing node discriminability by evaluating latent feature contributions. Extensive experiments demonstrate that our framework outperforms state-of-the-art graph SSL methods across benchmarks, achieving superior performance without contrastive loss or complex decoders. Key innovations include (1) a non-contrastive, view-invariant joint embedding predictive architecture, (2) Leveraging single context and multiple targets relationship between subgraphs, and (3) GMM-based pseudo-label scoring to capture semantic contributions. This work advances graph SSL by offering a computationally efficient, collapse-resistant paradigm that bridges spatial and semantic graph features for downstream tasks. The code for our paper can be found at https://github.com/Deceptrax123/JPEB-GSSL

comment: Preprint. Under Review

♻ ☆ ExMAG: Learning of Maximally Ancestral Graphs

Petr Ryšavý, Pavel Rytíř, Xiaoyu He, Georgios Korpas, Jakub Mareček

As one transitions from statistical to causal learning, one is seeking the most appropriate causal model. Dynamic Bayesian networks are a popular model, where a weighted directed acyclic graph represents the causal relationships. Stochastic processes are represented by its vertices, and weighted oriented edges suggest the strength of the causal relationships. When there are confounders, one would like to utilize both oriented edges (when the direction of causality is clear) and edges that are not oriented (when there is a confounder or not a relationship), yielding mixed graphs. A little-studied extension of acyclicity to this mixed-graph setting is known as maximally ancestral graphs with consideration of confounders. We propose a score-based learning algorithm for learning maximally ancestral graphs. A mixed-integer quadratic program is formulated, and an algorithmic approach is proposed, in which the pre-generation of exponentially many constraints is avoided by generating only violated constraints in the so-called branch-and-cut (``lazy constraint'') method. Comparing the novel approach to the state-of-the-art, we show that the proposed approach turns out to produce more accurate results when applied to small and medium-sized synthetic instances containing up to 25 variables.

♻ ☆ MTL-LoRA: Low-Rank Adaptation for Multi-Task Learning

Yaming Yang, Dilxat Muhtar, Yelong Shen, Yuefeng Zhan, Jianfeng Liu, Yujing Wang, Hao Sun, Denvy Deng, Feng Sun, Qi Zhang, Weizhu Chen, Yunhai Tong

Parameter-efficient fine-tuning (PEFT) has been widely employed for domain adaptation, with LoRA being one of the most prominent methods due to its simplicity and effectiveness. However, in multi-task learning (MTL) scenarios, LoRA tends to obscure the distinction between tasks by projecting sparse high-dimensional features from different tasks into the same dense low-dimensional intrinsic space. This leads to task interference and suboptimal performance for LoRA and its variants. To tackle this challenge, we propose MTL-LoRA, which retains the advantages of low-rank adaptation while significantly enhancing MTL capabilities. MTL-LoRA augments LoRA by incorporating additional task-adaptive parameters that differentiate task-specific information and capture shared knowledge across various tasks within low-dimensional spaces. This approach enables pre-trained models to jointly adapt to different target domains with a limited number of trainable parameters. Comprehensive experimental results, including evaluations on public academic benchmarks for natural language understanding, commonsense reasoning, and image-text understanding, as well as real-world industrial text Ads relevance datasets, demonstrate that MTL-LoRA outperforms LoRA and its various variants with comparable or even fewer learnable parameters in MTL setting.

comment: 12 Pages, 4 Figures

♻ ☆ FedECA: A Federated External Control Arm Method for Causal Inference with Time-To-Event Data in Distributed Settings

Jean Ogier du Terrail, Quentin Klopfenstein, Honghao Li, Imke Mayer, Nicolas Loiseau, Mohammad Hallal, Michael Debouver, Thibault Camalon, Thibault Fouqueray, Jorge Arellano Castro, Zahia Yanes, Laetitia Dahan, Julien Taïeb, Pierre Laurent-Puig, Jean-Baptiste Bachet, Shulin Zhao, Remy Nicolle, Jérome Cros, Daniel Gonzalez, Robert Carreras-Torres, Adelaida Garcia Velasco, Kawther Abdilleh, Sudheer Doss, Félix Balazard, Mathieu Andreux

External control arms (ECA) can inform the early clinical development of experimental drugs and provide efficacy evidence for regulatory approval. However, the main challenge in implementing ECA lies in accessing real-world or historical clinical trials data. Indeed, regulations protecting patients' rights by strictly controlling data processing make pooling data from multiple sources in a central server often difficult. To address these limitations, we develop a new method, 'FedECA' that leverages federated learning (FL) to enable inverse probability of treatment weighting (IPTW) for time-to-event outcomes on separate cohorts without needing to pool data. To showcase the potential of FedECA, we apply it in different settings of increasing complexity culminating with a real-world use-case in which FedECA is used to compare the treatment effect of two approved chemotherapy regimens using data from three separate cohorts of patients with metastatic pancreatic cancer. By sharing our code, we hope FedECA will foster the creation of federated research networks and thus accelerate drug development.

comment: code available at: https://github.com/owkin/fedeca, bug in SMD computation present in v1 and v2 fixed, many experiments on real data added + fix in YODA experiments using imputed data instead of raw data (v3->v4) + affiliations fix + more precise wording for acknowledgments, real-world experiment results fixed by excluding data with bias + text polished (v5->v6) + updating abstract(v6->v7)

♻ ☆ When Counterfactual Reasoning Fails: Chaos and Real-World Complexity

Yahya Aalaila, Gerrit Großmann, Sumantrak Mukherjee, Jonas Wahl, Sebastian Vollmer

Counterfactual reasoning, a cornerstone of human cognition and decision-making, is often seen as the 'holy grail' of causal learning, with applications ranging from interpreting machine learning models to promoting algorithmic fairness. While counterfactual reasoning has been extensively studied in contexts where the underlying causal model is well-defined, real-world causal modeling is often hindered by model and parameter uncertainty, observational noise, and chaotic behavior. The reliability of counterfactual analysis in such settings remains largely unexplored. In this work, we investigate the limitations of counterfactual reasoning within the framework of Structural Causal Models. Specifically, we empirically investigate \emph{counterfactual sequence estimation} and highlight cases where it becomes increasingly unreliable. We find that realistic assumptions, such as low degrees of model uncertainty or chaotic dynamics, can result in counterintuitive outcomes, including dramatic deviations between predicted and true counterfactual trajectories. This work urges caution when applying counterfactual reasoning in settings characterized by chaos and uncertainty. Furthermore, it raises the question of whether certain systems may pose fundamental limitations on the ability to answer counterfactual questions about their behavior.

♻ ☆ Lie Detector: Unified Backdoor Detection via Cross-Examination Framework

Xuan Wang, Siyuan Liang, Dongping Liao, Han Fang, Aishan Liu, Xiaochun Cao, Yu-liang Lu, Ee-Chien Chang, Xitong Gao

Institutions with limited data and computing resources often outsource model training to third-party providers in a semi-honest setting, assuming adherence to prescribed training protocols with pre-defined learning paradigm (e.g., supervised or semi-supervised learning). However, this practice can introduce severe security risks, as adversaries may poison the training data to embed backdoors into the resulting model. Existing detection approaches predominantly rely on statistical analyses, which often fail to maintain universally accurate detection accuracy across different learning paradigms. To address this challenge, we propose a unified backdoor detection framework in the semi-honest setting that exploits cross-examination of model inconsistencies between two independent service providers. Specifically, we integrate central kernel alignment to enable robust feature similarity measurements across different model architectures and learning paradigms, thereby facilitating precise recovery and identification of backdoor triggers. We further introduce backdoor fine-tuning sensitivity analysis to distinguish backdoor triggers from adversarial perturbations, substantially reducing false positives. Extensive experiments demonstrate that our method achieves superior detection performance, improving accuracy by 5.4%, 1.6%, and 11.9% over SoTA baselines across supervised, semi-supervised, and autoregressive learning tasks, respectively. Notably, it is the first to effectively detect backdoors in multimodal large language models, further highlighting its broad applicability and advancing secure deep learning.

♻ ☆ FedORGP: Guiding Heterogeneous Federated Learning with Orthogonality Regularization on Global Prototypes

Fucheng Guo, Zeyu Luan, Qing Li, Dan Zhao, Yong Jiang

Federated Learning (FL) has emerged as an essential framework for distributed machine learning, especially with its potential for privacy-preserving data processing. However, existing FL frameworks struggle to address statistical and model heterogeneity, which severely impacts model performance. While Heterogeneous Federated Learning (HtFL) introduces prototype-based strategies to address the challenges, current approaches face limitations in achieving optimal separation of prototypes. This paper presents FedORGP, a novel HtFL algorithm designed to improve global prototype separation through orthogonality regularization, which not only encourages intra-class prototype similarity but also significantly expands the inter-class angular separation. With the guidance of the global prototype, each client keeps its embeddings aligned with the corresponding prototype in the feature space, promoting directional independence that integrates seamlessly with the cross-entropy (CE) loss. We provide theoretical proof of FedORGP's convergence under non-convex conditions. Extensive experiments demonstrate that FedORGP outperforms seven state-of-the-art baselines, achieving up to 10.12\% accuracy improvement in scenarios where statistical and model heterogeneity coexist.

♻ ☆ $p$-Adic Polynomial Regression as Alternative to Neural Network for Approximating $p$-Adic Functions of Many Variables

Alexander P. Zubarev

A method for approximating continuous functions $\mathbb{Z}_{p}^{n}\rightarrow\mathbb{Z}_{p}$ by a linear superposition of continuous functions $\mathbb{Z}_{p}\rightarrow\mathbb{Z}_{p}$ is presented and a polynomial regression model is constructed that allows approximating such functions with any degree of accuracy. A physical interpretation of such a model is given and possible methods for its training are discussed. The proposed model can be considered as a simple alternative to possible $p$-adic models based on neural network architecture.

comment: 10 pages

♻ ☆ MambaPEFT: Exploring Parameter-Efficient Fine-Tuning for Mamba ICLR2025

Masakazu Yoshimura, Teruaki Hayashi, Yota Maeda

An ecosystem of Transformer-based models has been established by building large models with extensive data. Parameter-efficient fine-tuning (PEFT) is a crucial technology for deploying these models to downstream tasks with minimal cost while achieving effective performance. Recently, Mamba, a State Space Model (SSM)-based model, has attracted attention as a potential alternative to Transformers. While many large-scale Mamba-based models have been proposed, efficiently adapting pre-trained Mamba-based models to downstream tasks remains unexplored. In this paper, we conduct an exploratory analysis of PEFT methods for Mamba. We investigate the effectiveness of existing PEFT methods for Transformers when applied to Mamba. We also modify these methods to better align with the Mamba architecture. Additionally, we propose new Mamba-specific PEFT methods that leverage the distinctive structure of Mamba. Our experiments indicate that PEFT performs more effectively for Mamba than Transformers. Lastly, we demonstrate how to effectively combine multiple PEFT methods and provide a framework that outperforms previous works. To ensure reproducibility, we will release the code after publication.

comment: Accepted to ICLR2025

♻ ☆ Prior Learning in Introspective VAEs

Ioannis Athanasiadis, Fredrik Lindsten, Michael Felsberg

Variational Autoencoders (VAEs) are a popular framework for unsupervised learning and data generation. A plethora of methods have been proposed focusing on improving VAEs, with the incorporation of adversarial objectives and the integration of prior learning mechanisms being prominent directions. When it comes to the former, an indicative instance is the recently introduced family of Introspective VAEs aiming at ensuring that a low likelihood is assigned to unrealistic samples. In this study, we focus on the Soft-IntroVAE (S-IntroVAE) and investigate the implication of incorporating a multimodal and learnable prior into this framework. Namely, we formulate the prior as a third player and show that when trained in cooperation with the decoder constitutes an effective way for prior learning, which shares the Nash Equilibrium with the vanilla S-IntroVAE. Furthermore, based on a modified formulation of the optimal ELBO in S-IntroVAE, we develop theoretically motivated regularizations, that is (i) adaptive variance clipping to stabilize training when learning the prior and (ii) responsibility regularization to discourage the formation of inactive prior mode. Finally, we perform a series of targeted experiments on a 2D density estimation benchmark and in an image generation setting comprised of the (F)-MNIST and CIFAR-10 datasets demonstrating the benefit of prior learning in S-IntroVAE in generation and representation learning.

♻ ☆ AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors ICLR 2025

Ruoxuan Feng, Jiangyu Hu, Wenke Xia, Tianci Gao, Ao Shen, Yuhao Sun, Bin Fang, Di Hu

Visuo-tactile sensors aim to emulate human tactile perception, enabling robots to precisely understand and manipulate objects. Over time, numerous meticulously designed visuo-tactile sensors have been integrated into robotic systems, aiding in completing various tasks. However, the distinct data characteristics of these low-standardized visuo-tactile sensors hinder the establishment of a powerful tactile perception system. We consider that the key to addressing this issue lies in learning unified multi-sensor representations, thereby integrating the sensors and promoting tactile knowledge transfer between them. To achieve unified representation of this nature, we introduce TacQuad, an aligned multi-modal multi-sensor tactile dataset from four different visuo-tactile sensors, which enables the explicit integration of various sensors. Recognizing that humans perceive the physical environment by acquiring diverse tactile information such as texture and pressure changes, we further propose to learn unified multi-sensor representations from both static and dynamic perspectives. By integrating tactile images and videos, we present AnyTouch, a unified static-dynamic multi-sensor representation learning framework with a multi-level structure, aimed at both enhancing comprehensive perceptual abilities and enabling effective cross-sensor transfer. This multi-level architecture captures pixel-level details from tactile data via masked modeling and enhances perception and transferability by learning semantic-level sensor-agnostic features through multi-modal alignment and cross-sensor matching. We provide a comprehensive analysis of multi-sensor transferability, and validate our method on various datasets and in the real-world pouring task. Experimental results show that our method outperforms existing methods, exhibits outstanding static and dynamic perception capabilities across various sensors.

comment: Accepted by ICLR 2025

♻ ☆ A Clustering Method with Graph Maximum Decoding Information IJCNN 2024

Xinrun Xu, Manying Lv, Zhanbiao Lian, Yurong Wu, Jin Yan, Shan Jiang, Zhiming Ding

The clustering method based on graph models has garnered increased attention for its widespread applicability across various knowledge domains. Its adaptability to integrate seamlessly with other relevant applications endows the graph model-based clustering analysis with the ability to robustly extract "natural associations" or "graph structures" within datasets, facilitating the modelling of relationships between data points. Despite its efficacy, the current clustering method utilizing the graph-based model overlooks the uncertainty associated with random walk access between nodes and the embedded structural information in the data. To address this gap, we present a novel Clustering method for Maximizing Decoding Information within graph-based models, named CMDI. CMDI innovatively incorporates two-dimensional structural information theory into the clustering process, consisting of two phases: graph structure extraction and graph vertex partitioning. Within CMDI, graph partitioning is reformulated as an abstract clustering problem, leveraging maximum decoding information to minimize uncertainty associated with random visits to vertices. Empirical evaluations on three real-world datasets demonstrate that CMDI outperforms classical baseline methods, exhibiting a superior decoding information ratio (DI-R). Furthermore, CMDI showcases heightened efficiency, particularly when considering prior knowledge (PK). These findings underscore the effectiveness of CMDI in enhancing decoding information quality and computational efficiency, positioning it as a valuable tool in graph-based clustering analyses.

comment: 9 pages, 9 figures, IJCNN 2024

♻ ☆ Evaluating machine learning models for predicting pesticides toxicity to honey bees

Jakub Adamczyk, Jakub Poziemski, Pawel Siedlecki

Small molecules play a critical role in the biomedical, environmental, and agrochemical domains, each with distinct physicochemical requirements and success criteria. Although biomedical research benefits from extensive datasets and established benchmarks, agrochemical data remain scarce, particularly with respect to species-specific toxicity. This work focuses on ApisTox, the most comprehensive dataset of experimentally validated chemical toxicity to the honey bee (Apis mellifera), an ecologically vital pollinator. We evaluate ApisTox using a diverse suite of machine learning approaches, including molecular fingerprints, graph kernels, and graph neural networks, as well as pretrained models. Comparative analysis with medicinal datasets from the MoleculeNet benchmark reveals that ApisTox represents a distinct chemical space. Performance degradation on non-medicinal datasets, such as ApisTox, demonstrates their limited generalizability of current state-of-the-art algorithms trained solely on biomedical data. Our study highlights the need for more diverse datasets and for targeted model development geared toward the agrochemical domain.

♻ ☆ ResNLS: An Improved Model for Stock Price Forecasting

Yuanzhe Jia, Ali Anaissi, Basem Suleiman

Stock prices forecasting has always been a challenging task. Although many research projects try to address the problem, few of them pay attention to the varying degrees of dependencies between stock prices. In this paper, we introduce a hybrid model that improves the prediction of stock prices by emphasizing the dependencies between adjacent stock prices. The proposed model, ResNLS, is mainly composed of two neural architectures, ResNet and LSTM. ResNet serves as a feature extractor to identify dependencies between stock prices, while LSTM analyzes the initial time series data with the combination of dependencies, which are considered as residuals. Our experiment reveals that when the closing price data for the previous 5 consecutive trading days is used as input, the performance of the model (ResNLS-5) is optimal compared to those with other inputs. Furthermore, ResNLS-5 demonstrates at least a 20% improvement over current state-of-the-art baselines. To verify whether ResNLS-5 can help clients effectively avoid risks and earn profits in the stock market, we construct a quantitative trading framework for back testing. The result shows that the trading strategy based on ResNLS-5 predictions can successfully mitigate losses during declining stock prices and generate profits in periods of rising stock prices. The relevant code is publicly available on GitHub.

comment: Accepted by Computational Intelligence 2023

♻ ☆ Holistic analysis on the sustainability of Federated Learning across AI product lifecycle

Hongliu Cao

In light of emerging legal requirements and policies focused on privacy protection, there is a growing trend of companies across various industries adopting Federated Learning (FL). This decentralized approach involves multiple clients or silos, collaboratively training a global model under the coordination of a central server while utilizing their private local data. Unlike traditional methods that necessitate data sharing and transmission, Cross-Silo FL allows clients to share model updates rather than raw data, thereby enhancing privacy. Despite its growing adoption, the carbon impact associated with Cross-Silo FL remains poorly understood due to the limited research in this area. This study seeks to bridge this gap by evaluating the sustainability of Cross-Silo FL throughout the entire AI product lifecycle, extending the analysis beyond the model training phase alone. We systematically compare this decentralized method with traditional centralized approaches and present a robust quantitative framework for assessing the costs and CO2 emissions in real-world Cross-Silo FL environments. Our findings indicate that the energy consumption and costs of model training are comparable between Cross-Silo Federated Learning and Centralized Learning. However, the additional data transfer and storage requirements inherent in Centralized Learning can result in significant, often overlooked CO2 emissions. Moreover, we introduce an innovative data and application management system that integrates Cross-Silo FL and analytics, aiming at improving the sustainability and economic efficiency of IT enterprises.

comment: Presented in Sophia Summit 2023

♻ ☆ Vision-Language Models for Acute Tuberculosis Diagnosis: A Multimodal Approach Combining Imaging and Clinical Data

Ananya Ganapthy, Praveen Shastry, Naveen Kumarasami, Anandakumar D, Keerthana R, Mounigasri M, Varshinipriya M, Kishore Prasath Venkatesh, Bargava Subramanian, Kalyan Sivasailam

Background: This study introduces a Vision-Language Model (VLM) leveraging SIGLIP and Gemma-3b architectures for automated acute tuberculosis (TB) screening. By integrating chest X-ray images and clinical notes, the model aims to enhance diagnostic accuracy and efficiency, particularly in resource-limited settings. Methods: The VLM combines visual data from chest X-rays with clinical context to generate detailed, context-aware diagnostic reports. The architecture employs SIGLIP for visual encoding and Gemma-3b for decoding, ensuring effective representation of acute TB-specific pathologies and clinical insights. Results: Key acute TB pathologies, including consolidation, cavities, and nodules, were detected with high precision (97percent) and recall (96percent). The model demonstrated strong spatial localization capabilities and robustness in distinguishing TB-positive cases, making it a reliable tool for acute TB diagnosis. Conclusion: The multimodal capability of the VLM reduces reliance on radiologists, providing a scalable solution for acute TB screening. Future work will focus on improving the detection of subtle pathologies and addressing dataset biases to enhance its generalizability and application in diverse global healthcare settings.

comment: 11 pages, 3 figures

♻ ☆ 1-2-3-Go! Policy Synthesis for Parameterized Markov Decision Processes via Decision-Tree Learning and Generalization

Muqsit Azeem, Debraj Chakraborty, Sudeep Kanav, Jan Kretinsky, Mohammadsadegh Mohagheghi, Stefanie Mohr, Maximilian Weininger

Despite the advances in probabilistic model checking, the scalability of the verification methods remains limited. In particular, the state space often becomes extremely large when instantiating parameterized Markov decision processes (MDPs) even with moderate values. Synthesizing policies for such \emph{huge} MDPs is beyond the reach of available tools. We propose a learning-based approach to obtain a reasonable policy for such huge MDPs. The idea is to generalize optimal policies obtained by model-checking small instances to larger ones using decision-tree learning. Consequently, our method bypasses the need for explicit state-space exploration of large models, providing a practical solution to the state-space explosion problem. We demonstrate the efficacy of our approach by performing extensive experimentation on the relevant models from the quantitative verification benchmark set. The experimental results indicate that our policies perform well, even when the size of the model is orders of magnitude beyond the reach of state-of-the-art analysis tools.

comment: Extended version of the paper accepted at VMCAI 2025

♻ ☆ Designing Heterogeneous GNNs with Desired Permutation Properties for Wireless Resource Allocation

Jianyu Zhao, Chenyang Yang, Tingting Liu

Graph neural networks (GNNs) have been designed for learning a variety of wireless policies, i.e., the mappings from environment parameters to decision variables, thanks to their superior performance, and the potential in enabling scalability and size generalizability. These merits are rooted in leveraging permutation prior, i.e., satisfying the permutation property of the policy to be learned (referred to as desired permutation property). Many wireless policies are with complicated permutation properties. To satisfy these properties, heterogeneous GNNs (HetGNNs) should be used to learn such policies. There are two critical factors that enable a HetGNN to satisfy a desired permutation property: constructing an appropriate heterogeneous graph and judiciously designing the architecture of the HetGNN. However, both the graph and the HetGNN are designed heuristically so far. In this paper, we strive to provide a systematic approach for the design to satisfy the desired permutation property. We first propose a method for constructing a graph for a policy, where the edges and their types are defined for the sake of satisfying complicated permutation properties. Then, we provide and prove three sufficient conditions to design a HetGNN such that it can satisfy the desired permutation property when learning over an appropriate graph. These conditions suggest a method of designing the HetGNN with desired permutation property by sharing the processing, combining, and pooling functions according to the types of vertices and edges of the graph. We take power allocation and hybrid precoding policies as examples for demonstrating how to apply the proposed methods and validating the impact of the permutation prior by simulations.

♻ ☆ ZETA: Leveraging Z-order Curves for Efficient Top-k Attention ICLR

Qiuhao Zeng, Jerry Huang, Peng Lu, Gezheng Xu, Boxing Chen, Charles Ling, Boyu Wang

Over recent years, the Transformer has become a fundamental building block for sequence modeling architectures. Yet at its core is the use of self-attention, whose memory and computational cost grow quadratically with the sequence length $N$, rendering it prohibitively expensive for long sequences. A promising approach is top-$k$ attention, which selects only the $k$ most relevant tokens and achieves performance comparable to vanilla self-attention while significantly reducing space and computational demands. However, causal masks require the current query token to only attend to past tokens, preventing the existing top-$k$ attention method from efficiently searching for the most relevant tokens in parallel, thereby limiting training efficiency. In this work, we propose ZETA, leveraging \textbf{Z}-Order Curves for \textbf{E}fficient \textbf{T}op-$k$ \textbf{A}ttention, to enable parallel querying of past tokens for entire sequences. % in both space and time complexity of $\mathcal{O}(N \log N)$. We first theoretically show that the choice of key and query dimensions involves a trade-off between the curse of dimensionality and the preservation of relative distances after projection. In light of this insight, we propose reducing the dimensionality of keys and queries in contrast to values and further leverage $Z$-order curves to map low-dimensional keys and queries into \emph{one}-dimensional space, which permits parallel sorting, thereby largely improving the efficiency for top-$k$ token selection. Experimental results demonstrate that ZETA matches the performance of standard attention on the synthetic \textsc{Multi-Query Associative Recall} task and outperforms attention and its variants on \textsc{Long Range Arena} and \textsc{WikiText-103} language modeling.

comment: 25 pages, 4 figures, accepted in International Conference on Learning Representations (ICLR) 2025

♻ ☆ A predictive machine learning force field framework for liquid electrolyte development

Sheng Gong, Yumin Zhang, Zhenliang Mu, Zhichen Pu, Hongyi Wang, Zhiao Yu, Mengyi Chen, Tianze Zheng, Zhi Wang, Lifei Chen, Zhenze Yang, Xiaojie Wu, Shaochen Shi, Weihao Gao, Wen Yan, Liang Xiang

Despite the widespread applications of machine learning force fields (MLFF) in solids and small molecules, there is a notable gap in applying MLFF to simulate liquid electrolyte, a critical component of the current commercial lithium-ion battery. In this work, we introduce BAMBOO (\textbf{B}yteDance \textbf{A}I \textbf{M}olecular Simulation \textbf{Boo}ster), a predictive framework for molecular dynamics (MD) simulations, with a demonstration of its capability in the context of liquid electrolyte for lithium batteries. We design a physics-inspired graph equivariant transformer architecture as the backbone of BAMBOO to learn from quantum mechanical simulations. Additionally, we introduce an ensemble knowledge distillation approach and apply it to MLFFs to reduce the fluctuation of observations from MD simulations. Finally, we propose a density alignment algorithm to align BAMBOO with experimental measurements. BAMBOO demonstrates state-of-the-art accuracy in predicting key electrolyte properties such as density, viscosity, and ionic conductivity across various solvents and salt combinations. The current model, trained on more than 15 chemical species, achieves the average density error of 0.01 g/cm$^3$ on various compositions compared with experiment.

comment: Figures provided as the tex source files

♻ ☆ Buyer-Initiated Auction Mechanism for Data Redemption in Machine Unlearning

Bin Han, Di Feng, Jie Wang, Hans D. Schotten

The rapid growth of artificial intelligence (AI) has raised privacy concerns over user data, leading to regulations like the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). With the essential toolbox provided by machine unlearning, AI service providers are now able to remove user data from their trained models as well as the training datasets, so as to comply with such regulations. However, extensive data redemption can be costly and degrade model accuracy. To balance the cost of unlearning and the privacy protection, we propose a buyer-initiated auction mechanism for data redemption, enabling the service provider to purchase data from willing users with appropriate compensation. This approach does not require the server to have any a priori knowledge about the users' privacy preference, and provides an efficient solution for maximizing the social welfare in the investigated problem.

comment: Submitted to IEEE GLOBECOM 2025

♻ ☆ Sparse identification of nonlinear dynamics and Koopman operators with Shallow Recurrent Decoder Networks

Mars Liyao Gao, Jan P. Williams, J. Nathan Kutz

Modeling real-world spatio-temporal data is exceptionally difficult due to inherent high dimensionality, measurement noise, partial observations, and often expensive data collection procedures. In this paper, we present Sparse Identification of Nonlinear Dynamics with SHallow REcurrent Decoder networks (SINDy-SHRED), a method to jointly solve the sensing and model identification problems with simple implementation, efficient computation, and robust performance. SINDy-SHRED uses Gated Recurrent Units to model the temporal sequence of sparse sensor measurements along with a shallow decoder network to reconstruct the full spatio-temporal field from the latent state space. Our algorithm introduces a SINDy-based regularization for which the latent space progressively converges to a SINDy-class functional, provided the projection remains within the set. In restricting SINDy to a linear model, a Koopman-SHRED model is generated. SINDy-SHRED (i) learns a symbolic and interpretable generative model of a parsimonious and low-dimensional latent space for the complex spatio-temporal dynamics, (ii) discovers new physics models even for well-known physical systems, (iii) achieves provably robust convergence with an observed globally convex loss landscape, and (iv) achieves superior accuracy, data efficiency, and training time, all with fewer model parameters. We conduct systematic experimental studies on PDE data such as turbulent flows, real-world sensor measurements for sea surface temperature, and direct video data. The interpretable SINDy and Koopman models of latent state dynamics enable stable and accurate long-term video predictions, outperforming all current baseline deep learning models in accuracy, training time, and data requirements, including Convolutional LSTM, PredRNN, ResNet, and SimVP.

♻ ☆ Diffusion State-Guided Projected Gradient for Inverse Problems ICLR 2025

Rayhan Zirvi, Bahareh Tolooshams, Anima Anandkumar

Recent advancements in diffusion models have been effective in learning data priors for solving inverse problems. They leverage diffusion sampling steps for inducing a data prior while using a measurement guidance gradient at each step to impose data consistency. For general inverse problems, approximations are needed when an unconditionally trained diffusion model is used since the measurement likelihood is intractable, leading to inaccurate posterior sampling. In other words, due to their approximations, these methods fail to preserve the generation process on the data manifold defined by the diffusion prior, leading to artifacts in applications such as image restoration. To enhance the performance and robustness of diffusion models in solving inverse problems, we propose Diffusion State-Guided Projected Gradient (DiffStateGrad), which projects the measurement gradient onto a subspace that is a low-rank approximation of an intermediate state of the diffusion process. DiffStateGrad, as a module, can be added to a wide range of diffusion-based inverse solvers to improve the preservation of the diffusion process on the prior manifold and filter out artifact-inducing components. We highlight that DiffStateGrad improves the robustness of diffusion models in terms of the choice of measurement guidance step size and noise while improving the worst-case performance. Finally, we demonstrate that DiffStateGrad improves upon the state-of-the-art on linear and nonlinear image restoration inverse problems. Our code is available at https://github.com/Anima-Lab/DiffStateGrad.

comment: Published as a conference paper at ICLR 2025. RZ and BT have equal contributions

♻ ☆ CodingTeachLLM: Empowering LLM's Coding Ability via AST Prior Knowledge

Zhangquan Chen, Chunjiang Liu, Haobin Duan

In this paper, we introduce CodingTeachLLM, a large language model (LLM) designed for coding teaching. Specially, we aim to enhance the coding ability of LLM and lead it to better teaching mode in education context. Thus, we propose an end-to-end prior-based three-phases supervised fine-tuned model, which is proved more competitive than traditional fine-tuning method. More specifically, our model realizes the structural disassembly and incremental guided output of educational knowledge. To this end, we robustify data classification of three types via a sampler and overlap estimation neural network, and inject the preprocessing datasets into pre-trained model in three batches for LORA fine-tuning. Then, we design a prior module couples system prompt, vector databases, and abstract syntax tree task segmentation. Finally, the compression method and regularization constraint are applied to the prior-based fine-tuned model, followed by text filter at the output end to obtain incremental guided results. Our model represents the first research effort to truly embody the tutor role with the features of abundant educational knowledge, step-by-step incremental guided outputs and non-disclosure of answers. Extensive experiments report that our model also achieves state-of-the-art in code abilities compared to open-source models, reaching an impressive 75.10% on the HumanEval (@pass 1) benchmark. Additionally, our model maintains strong conversational capabilities, with the 13B quantized version achieving scores of 56.34, 50.60, and 45.27 respectively on the MMLU, C-Eval, and AGIEval (5 shot) dialogue evaluation benchmarks.

comment: 9 pages, 2 figures

♻ ☆ Temporal and Semantic Evaluation Metrics for Foundation Models in Post-Hoc Analysis of Robotic Sub-tasks IROS 2024

Jonathan Salfity, Selma Wanna, Minkyu Choi, Mitch Pryor

Recent works in Task and Motion Planning (TAMP) show that training control policies on language-supervised robot trajectories with quality labeled data markedly improves agent task success rates. However, the scarcity of such data presents a significant hurdle to extending these methods to general use cases. To address this concern, we present an automated framework to decompose trajectory data into temporally bounded and natural language-based descriptive sub-tasks by leveraging recent prompting strategies for Foundation Models (FMs) including both Large Language Models (LLMs) and Vision Language Models (VLMs). Our framework provides both time-based and language-based descriptions for lower-level sub-tasks that comprise full trajectories. To rigorously evaluate the quality of our automatic labeling framework, we contribute an algorithm SIMILARITY to produce two novel metrics, temporal similarity and semantic similarity. The metrics measure the temporal alignment and semantic fidelity of language descriptions between two sub-task decompositions, namely an FM sub-task decomposition prediction and a ground-truth sub-task decomposition. We present scores for temporal similarity and semantic similarity above 90%, compared to 30% of a randomized baseline, for multiple robotic environments, demonstrating the effectiveness of our proposed framework. Our results enable building diverse, large-scale, language-supervised datasets for improved robotic TAMP.

comment: 8 pages, 3 figures. IROS 2024 Submission

♻ ☆ Time-Series Forecasting via Topological Information Supervised Framework with Efficient Topological Feature Learning

ZiXin Lin, Nur Fariha Syaqina Zulkepli

Topological Data Analysis (TDA) has emerged as a powerful tool for extracting meaningful features from complex data structures, driving significant advancements in fields such as neuroscience, biology, machine learning, and financial modeling. Despite its success, the integration of TDA with time-series prediction remains underexplored due to three primary challenges: the limited utilization of temporal dependencies within topological features, computational bottlenecks associated with persistent homology, and the deterministic nature of TDA pipelines restricting generalized feature learning. This study addresses these challenges by proposing the Topological Information Supervised (TIS) Prediction framework, which leverages neural networks and Conditional Generative Adversarial Networks (CGANs) to generate synthetic topological features, preserving their distribution while significantly reducing computational time. We propose a novel training strategy that integrates topological consistency loss to improve the predictive accuracy of deep learning models. Specifically, we introduce two state-of-the-art models, TIS-BiGRU and TIS-Informer, designed to capture short-term and long-term temporal dependencies, respectively. Comparative experimental results demonstrate the superior performance of TIS models over conventional predictors, validating the effectiveness of integrating topological information. This work not only advances TDA-based time-series prediction but also opens new avenues for utilizing topological features in deep learning architectures.

comment: The experiments are incomplete

♻ ☆ Data-Free Group-Wise Fully Quantized Winograd Convolution via Learnable Scales CVPR 2025

Shuokai Pan, Gerti Tuzi, Sudarshan Sreeram, Dibakar Gope

Despite the revolutionary breakthroughs of large-scale text-to-image diffusion models for complex vision and downstream tasks, their extremely high computational and storage costs limit their usability. Quantization of diffusion models has been explored in recent works to reduce compute costs and memory bandwidth usage. To further improve inference time, fast convolution algorithms such as Winograd can be used for convolution layers, which account for a significant portion of computations in diffusion models. However, the significant quality loss of fully quantized Winograd using existing coarser-grained post-training quantization methods, combined with the complexity and cost of finetuning the Winograd transformation matrices for such large models to recover quality, makes them unsuitable for large-scale foundation models. Motivated by the presence of a large range of values in them, we investigate the impact of finer-grained group-wise quantization in quantizing diffusion models. While group-wise quantization can largely handle the fully quantized Winograd convolution, it struggles to deal with the large distribution imbalance in a sizable portion of the Winograd domain computation. To reduce range differences in the Winograd domain, we propose finetuning only the scale parameters of the Winograd transform matrices without using any domain-specific training data. Because our method does not depend on any training data, the generalization performance of quantized diffusion models is safely guaranteed. For text-to-image generation task, the 8-bit fully-quantized diffusion model with Winograd provides near-lossless quality (FID and CLIP scores) in comparison to the full-precision model. For image classification, our method outperforms the state-of-the-art Winograd PTQ method by 1.62% and 2.56% in top-1 ImageNet accuracy on ResNet18 and ResNet-34, respectively, with Winograd F(6, 3).

comment: Accepted by CVPR 2025

♻ ☆ DC-SGD: Differentially Private SGD with Dynamic Clipping through Gradient Norm Distribution Estimation

Chengkun Wei, Weixian Li, Chen Gong, Wenzhi Chen

Differentially Private Stochastic Gradient Descent (DP-SGD) is a widely adopted technique for privacy-preserving deep learning. A critical challenge in DP-SGD is selecting the optimal clipping threshold C, which involves balancing the trade-off between clipping bias and noise magnitude, incurring substantial privacy and computing overhead during hyperparameter tuning. In this paper, we propose Dynamic Clipping DP-SGD (DC-SGD), a framework that leverages differentially private histograms to estimate gradient norm distributions and dynamically adjust the clipping threshold C. Our framework includes two novel mechanisms: DC-SGD-P and DC-SGD-E. DC-SGD-P adjusts the clipping threshold based on a percentile of gradient norms, while DC-SGD-E minimizes the expected squared error of gradients to optimize C. These dynamic adjustments significantly reduce the burden of hyperparameter tuning C. The extensive experiments on various deep learning tasks, including image classification and natural language processing, show that our proposed dynamic algorithms achieve up to 9 times acceleration on hyperparameter tuning than DP-SGD. And DC-SGD-E can achieve an accuracy improvement of 10.62% on CIFAR10 than DP-SGD under the same privacy budget of hyperparameter tuning. We conduct rigorous theoretical privacy and convergence analyses, showing that our methods seamlessly integrate with the Adam optimizer. Our results highlight the robust performance and efficiency of DC-SGD, offering a practical solution for differentially private deep learning with reduced computational overhead and enhanced privacy guarantees.

comment: Accepted at IEEE Transactions on Information Forensics & Security

♻ ☆ Self-Vocabularizing Training for Neural Machine Translation NAACL

Pin-Jie Lin, Ernie Chang, Yangyang Shi, Vikas Chandra

Past vocabulary learning techniques identify relevant vocabulary before training, relying on statistical and entropy-based assumptions that largely neglect the role of model training. Empirically, we observe that trained translation models are induced to use a byte-pair encoding (BPE) vocabulary subset distinct from the original BPE vocabulary, leading to performance improvements when retrained with the induced vocabulary. In this paper, we analyze this discrepancy in neural machine translation by examining vocabulary and entropy shifts during self-training--where each iteration generates a labeled dataset by pairing source sentences with the model's predictions to define a new vocabulary. Building on these insights, we propose self-vocabularizing training, an iterative method that self-selects a smaller, more optimal vocabulary, yielding up to a 1.49 BLEU improvement. Moreover, we find that deeper model architectures lead to both an increase in unique token usage and a 6-8% reduction in vocabulary size.

comment: Accepted to NAACL SRW 2025

♻ ☆ Conditional Variable Flow Matching: Transforming Conditional Densities with Amortized Conditional Optimal Transport

Adam P. Generale, Andreas E. Robertson, Surya R. Kalidindi

Forecasting conditional stochastic nonlinear dynamical systems is a fundamental challenge repeatedly encountered across the biological and physical sciences. While flow-based models can impressively predict the temporal evolution of probability distributions representing possible outcomes of a specific process, existing frameworks cannot satisfactorily account for the impact of conditioning variables on these dynamics. Amongst several limitations, existing methods require training data with paired conditions and are developed for discrete conditioning variables. We propose Conditional Variable Flow Matching (CVFM), a framework for learning flows transforming conditional distributions with amortization across continuous conditioning variables - permitting predictions across the conditional density manifold. This is accomplished through several novel advances. In particular, simultaneous sample conditioned flows over the main and conditioning variables, alongside a conditional Wasserstein distance combined with a loss reweighting kernel facilitating conditional optimal transport. Collectively, these advances allow for learning system dynamics provided measurement data whose states and conditioning variables are not in correspondence. We demonstrate CVFM on a suite of increasingly challenging problems, including discrete and continuous conditional mapping benchmarks, image-to-image domain transfer, and modeling the temporal evolution of materials internal structure during manufacturing processes. We observe that CVFM results in improved performance and convergence characteristics over alternative conditional variants.

♻ ☆ Lean Formalization of Generalization Error Bound by Rademacher Complexity

Sho Sonoda, Kazumi Kasaura, Yuma Mizuno, Kei Tsukamoto, Naoto Onda

We formalize the generalization error bound using Rademacher complexity in the Lean 4 theorem prover. Generalization error quantifies the gap between a learning machine's performance on given training data versus unseen test data, and Rademacher complexity serves as an estimate of this error based on the complexity of learning machines, or hypothesis class. Unlike traditional methods such as PAC learning and VC dimension, Rademacher complexity is applicable across diverse machine learning scenarios including deep learning and kernel methods. We formalize key concepts and theorems, including the empirical and population Rademacher complexities, and establish generalization error bounds through formal proofs of McDiarmid's inequality, Hoeffding's lemma, and symmetrization arguments.

comment: modified a typo in affiliation

♻ ☆ Assessing and improving reliability of neighbor embedding methods: a map-continuity perspective

Zhexuan Liu, Rong Ma, Yiqiao Zhong

Visualizing high-dimensional data is essential for understanding biomedical data and deep learning models. Neighbor embedding methods, such as t-SNE and UMAP, are widely used but can introduce misleading visual artifacts. We find that the manifold learning interpretations from many prior works are inaccurate and that the misuse stems from a lack of data-independent notions of embedding maps, which project high-dimensional data into a lower-dimensional space. Leveraging the leave-one-out principle, we introduce LOO-map, a framework that extends embedding maps beyond discrete points to the entire input space. We identify two forms of map discontinuity that distort visualizations: one exaggerates cluster separation and the other creates spurious local structures. As a remedy, we develop two types of point-wise diagnostic scores to detect unreliable embedding points and improve hyperparameter selection, which are validated on datasets from computer vision and single-cell omics.

comment: 49 pages, 20 figures

♻ ☆ Non-Determinism of "Deterministic" LLM Settings

Berk Atil, Sarp Aykent, Alexa Chittams, Lisheng Fu, Rebecca J. Passonneau, Evan Radcliffe, Guru Rajan Rajagopal, Adam Sloan, Tomasz Tudrej, Ferhan Ture, Zhe Wu, Lixinyu Xu, Breck Baldwin

LLM (large language model) practitioners commonly notice that outputs can vary for the same inputs under settings expected to be deterministic. Yet the questions of how pervasive this is, and with what impact on results, have not to our knowledge been systematically investigated. We investigate non-determinism in five LLMs configured to be deterministic when applied to eight common tasks in across 10 runs, in both zero-shot and few-shot settings. We see accuracy variations up to 15% across naturally occurring runs with a gap of best possible performance to worst possible performance up to 70%. In fact, none of the LLMs consistently delivers repeatable accuracy across all tasks, much less identical output strings. Sharing preliminary results with insiders has revealed that non-determinism perhaps essential to the efficient use of compute resources via co-mingled data in input buffers so this issue is not going away anytime soon. To better quantify our observations, we introduce metrics focused on quantifying determinism, TARr@N for the total agreement rate at N runs over raw output, and TARa@N for total agreement rate of parsed-out answers. Our code and data are publicly available at http://github.com/REDACTED.

♻ ☆ PharMolixFM: All-Atom Foundation Models for Molecular Modeling and Generation

Yizhen Luo, Jiashuo Wang, Siqi Fan, Zaiqing Nie

Structural biology relies on accurate three-dimensional biomolecular structures to advance our understanding of biological functions, disease mechanisms, and therapeutics. While recent advances in deep learning have enabled the development of all-atom foundation models for molecular modeling and generation, existing approaches face challenges in generalization due to the multi-modal nature of atomic data and the lack of comprehensive analysis of training and sampling strategies. To address these limitations, we propose PharMolixFM, a unified framework for constructing all-atom foundation models based on multi-modal generative techniques. Our framework includes three variants using state-of-the-art multi-modal generative models. By formulating molecular tasks as a generalized denoising process with task-specific priors, PharMolixFM achieves robust performance across various structural biology applications. Experimental results demonstrate that PharMolixFM-Diff achieves competitive prediction accuracy in protein-small-molecule docking (83.9% vs. 90.2% RMSD < 2{\AA}, given pocket) with significantly improved inference speed. Moreover, we explore the empirical inference scaling law by introducing more sampling repeats or steps. Our code and model are available at https://github.com/PharMolix/OpenBioMed.

♻ ☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems

Ooha Lakkadi Reddy

This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.

comment: 106 pages (42 main text, 6 references, 58 appendices). 21 figures, 4 tables in main text; 106 figures, 8 tables total. Code: https://github.com/oohalakkadi/ivc2tyc. Undergraduate thesis at Duke Kunshan University. Accepted for presentation at the 52nd International Conference for Computer Applications & Quantitative Methods in Archaeology (CAA 2025), Athens, Greece

♻ ☆ Multi-objective Combinatorial Methodology for Nuclear Reactor Site Assessment: A Case Study for the United States

Omer Erdem, Kevin Daley, Gabrielle Hoelzle, Majdi I. Radaideh

As clean energy demand grows to meet sustainability and net-zero goals, nuclear energy emerges as a reliable option. However, high capital costs remain a challenge for nuclear power plants (NPP), where repurposing coal power plant sites (CPP) with existing infrastructure is one way to reduce these costs. Additionally, Brownfield sites-previously developed or underutilized lands often impacted by industrial activity-present another compelling alternative. This study introduces a novel multi-objective optimization methodology, leveraging combinatorial search to evaluate over 30,000 potential NPP sites in the United States. Our approach addresses gaps in the current practice of assigning pre-determined weights to each site attribute that could lead to bias in the ranking. Each site is assigned a performance-based score, derived from a detailed combinatorial analysis of its site attributes. The methodology generates a comprehensive database comprising site locations (inputs), attributes (outputs), site score (outputs), and the contribution of each attribute to the site score. We then use this database to train a neural network model, enabling rapid predictions of nuclear siting suitability across any location in the United States. Our findings highlight that CPP sites are highly competitive for nuclear development, but some Brownfield sites are able to compete with them. Notably, four CPP sites in Ohio, North Carolina, and New Hampshire, and two Brownfield sites in Florida and California rank among the most promising locations. These results underscore the potential of integrating machine learning and optimization techniques to transform nuclear siting, paving the way for a cost-effective and sustainable energy future.

comment: 32 Pages, 7 Tables, 12 Figures

♻ ☆ UniFlow: A Foundation Model for Unified Urban Spatio-Temporal Flow Prediction

Yuan Yuan, Jingtao Ding, Chonghua Han, Zhi Sheng, Depeng Jin, Yong Li

Urban spatio-temporal flow prediction, encompassing traffic flows and crowd flows, is crucial for optimizing city infrastructure and managing traffic and emergency responses. Traditional approaches have relied on separate models tailored to either grid-based data, representing cities as uniform cells, or graph-based data, modeling cities as networks of nodes and edges. In this paper, we build UniFlow, a foundational model for general urban flow prediction that unifies both grid-based and graphbased data. We first design a multi-view spatio-temporal patching mechanism to standardize different data into a consistent sequential format and then introduce a spatio-temporal transformer architecture to capture complex correlations and dynamics. To leverage shared spatio-temporal patterns across different data types and facilitate effective cross-learning, we propose SpatioTemporal Memory Retrieval Augmentation (ST-MRA). By creating structured memory modules to store shared spatio-temporal patterns, ST-MRA enhances predictions through adaptive memory retrieval. Extensive experiments demonstrate that UniFlow outperforms existing models in both grid-based and graph-based flow prediction, excelling particularly in scenarios with limited data availability, showcasing its superior performance and broad applicability. The datasets and code implementation have been released on https://github.com/YuanYuan98/UniFlow.

♻ ☆ Individualized Policy Evaluation and Learning under Clustered Network Interference

Yi Zhang, Kosuke Imai

Although there is now a large literature on policy evaluation and learning, much of the prior work assumes that the treatment assignment of one unit does not affect the outcome of another unit. Unfortunately, ignoring interference can lead to biased policy evaluation and ineffective learned policies. For example, treating influential individuals who have many friends can generate positive spillover effects, thereby improving the overall performance of an individualized treatment rule (ITR). We consider the problem of evaluating and learning an optimal ITR under clustered network interference (also known as partial interference), where clusters of units are sampled from a population and units may influence one another within each cluster. Unlike previous methods that impose strong restrictions on spillover effects, such as anonymous interference, the proposed methodology only assumes a semiparametric structural model, where each unit's outcome is an additive function of individual treatments within the cluster. Under this model, we propose an estimator that can be used to evaluate the empirical performance of an ITR. We show that this estimator is substantially more efficient than the standard inverse probability weighting estimator, which does not impose any assumption about spillover effects. We derive the finite-sample regret bound for a learned ITR, showing that the use of our efficient evaluation estimator leads to the improved performance of learned policies. We consider both experimental and observational studies, and for the latter, we develop a doubly robust estimator that is semiparametrically efficient and yields an optimal regret bound. Finally, we conduct simulation and empirical studies to illustrate the advantages of the proposed methodology.

♻ ☆ Provably-Safe Neural Network Training Using Hybrid Zonotope Reachability Analysis

Long Kiu Chung, Shreyas Kousik

Even though neural networks are being increasingly deployed in safety-critical control applications, it remains difficult to enforce constraints on their output, meaning that it is hard to guarantee safety in such settings. While many existing methods seek to verify a neural network's satisfaction of safety constraints, few address how to correct an unsafe network. The handful of works that extract a training signal from verification cannot handle non-convex sets, and are either conservative or slow. To begin addressing these challenges, this work proposes a neural network training method that can encourage the exact image of a non-convex input set for a neural network with rectified linear unit (ReLU) nonlinearities to avoid a non-convex unsafe region. This is accomplished by reachability analysis with scaled hybrid zonotopes, a modification of the existing hybrid zonotope set representation that enables parameterized scaling of non-convex polytopic sets with a differentiable collision check via mixed-integer linear programs (MILPs). The proposed method was shown to be effective and fast for networks with up to 240 neurons, with the computational complexity dominated by inverse operations on matrices that scale linearly in size with the number of neurons and complexity of input and unsafe sets. We demonstrate the practicality of our method by training a forward-invariant neural network controller for a non-convex input set to an affine system, as well as generating safe reach-avoid plans for a black-box dynamical system.

comment: 8 pages, 3 figures

♻ ☆ Robust Bayesian Optimization via Localized Online Conformal Prediction

Dongwon Kim, Matteo Zecchin, Sangwoo Park, Joonhyuk Kang, Osvaldo Simeone

Bayesian optimization (BO) is a sequential approach for optimizing black-box objective functions using zeroth-order noisy observations. In BO, Gaussian processes (GPs) are employed as probabilistic surrogate models to estimate the objective function based on past observations, guiding the selection of future queries to maximize utility. However, the performance of BO heavily relies on the quality of these probabilistic estimates, which can deteriorate significantly under model misspecification. To address this issue, we introduce localized online conformal prediction-based Bayesian optimization (LOCBO), a BO algorithm that calibrates the GP model through localized online conformal prediction (CP). LOCBO corrects the GP likelihood based on predictive sets produced by LOCBO, and the corrected GP likelihood is then denoised to obtain a calibrated posterior distribution on the objective function. The likelihood calibration step leverages an input-dependent calibration threshold to tailor coverage guarantees to different regions of the input space. Under minimal noise assumptions, we provide theoretical performance guarantees for LOCBO's iterates that hold for the unobserved objective function. These theoretical findings are validated through experiments on synthetic and real-world optimization tasks, demonstrating that LOCBO consistently outperforms state-of-the-art BO algorithms in the presence of model misspecification.

♻ ☆ Convolutional Neural Networks Can (Meta-)Learn the Same-Different Relation

Max Gupta, Sunayana Rane, R. Thomas McCoy, Thomas L. Griffiths

While convolutional neural networks (CNNs) have come to match and exceed human performance in many settings, the tasks these models optimize for are largely constrained to the level of individual objects, such as classification and captioning. Humans remain vastly superior to CNNs in visual tasks involving relations, including the ability to identify two objects as `same' or `different'. A number of studies have shown that while CNNs can be coaxed into learning the same-different relation in some settings, they tend to generalize poorly to other instances of this relation. In this work we show that the same CNN architectures that fail to generalize the same-different relation with conventional training are able to succeed when trained via meta-learning, which explicitly encourages abstraction and generalization across tasks.

♻ ☆ Enhancing Domain Adaptation through Prompt Gradient Alignment NeurIPS 2024

Hoang Phan, Lam Tran, Quyen Tran, Trung Le

Prior Unsupervised Domain Adaptation (UDA) methods often aim to train a domain-invariant feature extractor, which may hinder the model from learning sufficiently discriminative features. To tackle this, a line of works based on prompt learning leverages the power of large-scale pre-trained vision-language models to learn both domain-invariant and specific features through a set of domain-agnostic and domain-specific learnable prompts. Those studies typically enforce invariant constraints on representation, output, or prompt space to learn such prompts. In contrast, we cast UDA as a multiple-objective optimization problem in which each objective is represented by a domain loss. Under this new framework, we propose to align per-objective gradients to foster consensus between them. Additionally, to prevent potential overfitting when fine-tuning this deep learning architecture, we penalize the norm of these gradients. To achieve these goals, we devise a practical gradient update procedure that can work under both single-source and multi-source UDA. Empirically, our method consistently outperforms other vision-language model adaptation methods. The implementation is available at https://github.com/VietHoang1512/PGA.

comment: Accepted to NeurIPS 2024

♻ ☆ Causal Dynamic Variational Autoencoder for Counterfactual Regression in Longitudinal Data

Mouad El Bouchattaoui, Myriam Tami, Benoit Lepetit, Paul-Henry Cournède

Estimating treatment effects over time is relevant in many real-world applications, such as precision medicine, epidemiology, economy, and marketing. Many state-of-the-art methods either assume the observations of all confounders or seek to infer the unobserved ones. We take a different perspective by assuming unobserved risk factors, i.e., adjustment variables that affect only the sequence of outcomes. Under unconfoundedness, we target the Individual Treatment Effect (ITE) estimation with unobserved heterogeneity in the treatment response due to missing risk factors. We address the challenges posed by time-varying effects and unobserved adjustment variables. Led by theoretical results over the validity of the learned adjustment variables and generalization bounds over the treatment effect, we devise Causal DVAE (CDVAE). This model combines a Dynamic Variational Autoencoder (DVAE) framework with a weighting strategy using propensity scores to estimate counterfactual responses. The CDVAE model allows for accurate estimation of ITE and captures the underlying heterogeneity in longitudinal data. Evaluations of our model show superior performance over state-of-the-art models.

♻ ☆ Independent and Decentralized Learning in Markov Potential Games

Chinmay Maheshwari, Manxi Wu, Druv Pai, Shankar Sastry

We study a multi-agent reinforcement learning dynamics, and analyze its asymptotic behavior in infinite-horizon discounted Markov potential games. We focus on the independent and decentralized setting, where players do not know the game parameters, and cannot communicate or coordinate. In each stage, players update their estimate of Q-function that evaluates their total contingent payoff based on the realized one-stage reward in an asynchronous manner. Then, players independently update their policies by incorporating an optimal one-stage deviation strategy based on the estimated Q-function. Inspired by the actor-critic algorithm in single-agent reinforcement learning, a key feature of our learning dynamics is that agents update their Q-function estimates at a faster timescale than the policies. Leveraging tools from two-timescale asynchronous stochastic approximation theory, we characterize the convergent set of learning dynamics.

comment: 43 pages, 1 figure

♻ ☆ Addressing pitfalls in implicit unobserved confounding synthesis using explicit block hierarchical ancestral sampling

Xudong Sun, Alex Markham, Pratik Misra, Carsten Marr

Unbiased data synthesis is crucial for evaluating causal discovery algorithms in the presence of unobserved confounding, given the scarcity of real-world datasets. A common approach, implicit parameterization, encodes unobserved confounding by modifying the off-diagonal entries of the idiosyncratic covariance matrix while preserving positive definiteness. Within this approach, we identify that state-of-the-art protocols have two distinct issues that hinder unbiased sampling from the complete space of causal models: first, we give a detailed analysis of use of diagonally dominant constructions restricts the spectrum of partial correlation matrices; and second, the restriction of possible graphical structures when sampling bidirected edges, unnecessarily ruling out valid causal models. To address these limitations, we propose an improved explicit modeling approach for unobserved confounding, leveraging block-hierarchical ancestral generation of ground truth causal graphs. Algorithms for converting the ground truth DAG into ancestral graph is provided so that the output of causal discovery algorithms could be compared with. We draw connections between implicit and explicit parameterization, prove that our approach fully covers the space of causal models, including those generated by the implicit parameterization, thus enabling more robust evaluation of methods for causal discovery and inference.

♻ ☆ Disentangling Safe and Unsafe Corruptions via Anisotropy and Locality

Ramchandran Muthukumar, Ambar Pal, Jeremias Sulam, Rene Vidal

State-of-the-art machine learning systems are vulnerable to small perturbations to their input, where ``small'' is defined according to a threat model that assigns a positive threat to each perturbation. Most prior works define a task-agnostic, isotropic, and global threat, like the $\ell_p$ norm, where the magnitude of the perturbation fully determines the degree of the threat and neither the direction of the attack nor its position in space matter. However, common corruptions in computer vision, such as blur, compression, or occlusions, are not well captured by such threat models. This paper proposes a novel threat model called \texttt{Projected Displacement} (PD) to study robustness beyond existing isotropic and global threat models. The proposed threat model measures the threat of a perturbation via its alignment with \textit{unsafe directions}, defined as directions in the input space along which a perturbation of sufficient magnitude changes the ground truth class label. Unsafe directions are identified locally for each input based on observed training data. In this way, the PD threat model exhibits anisotropy and locality. Experiments on Imagenet-1k data indicate that, for any input, the set of perturbations with small PD threat includes \textit{safe} perturbations of large $\ell_p$ norm that preserve the true label, such as noise, blur and compression, while simultaneously excluding \textit{unsafe} perturbations that alter the true label. Unlike perceptual threat models based on embeddings of large-vision models, the PD threat model can be readily computed for arbitrary classification tasks without pre-training or finetuning. Further additional task annotation such as sensitivity to image regions or concept hierarchies can be easily integrated into the assessment of threat and thus the PD threat model presents practitioners with a flexible, task-driven threat specification.

comment: Published at IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025. Updated Acknowledgements

♻ ☆ Challenging Dataset and Multi-modal Gated Mixture of Experts Model for Remote Sensing Copy-Move Forgery Understanding

Ze Zhang, Enyuan Zhao, Yi Jiang, Jie Nie, Xinyue Liang

The Remote Sensing Copy-Move Question Answering (RSCMQA) task focuses on interpreting complex tampering scenarios and inferring the relationships between objects. Currently, publicly available datasets often use randomly generated tampered images, which lack spatial logic and do not meet the practical needs of defense security and land resource monitoring. To address this, we propose a high-quality manually annotated RSCMQA dataset, Real-RSCM, which provides more realistic evaluation metrics for the identification and understanding of remote sensing image tampering. The tampered images in the Real-RSCM dataset are subtle, authentic, and challenging, posing significant difficulties for model discrimination capabilities. To overcome these challenges, we introduce a multimodal gated mixture of experts model (CM-MMoE), which guides multi-expert models to discern tampered information in images through multi-level visual semantics and textual joint modeling. Extensive experiments demonstrate that CM-MMoE provides a stronger benchmark for the RSCMQA task compared to general VQA and CMQA models. Our dataset and code are available at https://github.com/shenyedepisa/CM-MMoE.

comment: 6 pages, 6 figures

♻ ☆ TeleAntiFraud-28k: An Audio-Text Slow-Thinking Dataset for Telecom Fraud Detection

Zhiming Ma, Peidong Wang, Minhua Huang, Jingpeng Wang, Kai Wu, Xiangzhao Lv, Yachun Pang, Yin Yang, Wenjie Tang, Yuchen Kang

The detection of telecom fraud faces significant challenges due to the lack of high-quality multimodal training data that integrates audio signals with reasoning-oriented textual analysis. To address this gap, we present TeleAntiFraud-28k, the first open-source audio-text slow-thinking dataset specifically designed for automated telecom fraud analysis. Our dataset is constructed through three strategies: (1) Privacy-preserved text-truth sample generation using automatically speech recognition (ASR)-transcribed call recordings (with anonymized original audio), ensuring real-world consistency through text-to-speech (TTS) model regeneration; (2) Semantic enhancement via large language model (LLM)-based self-instruction sampling on authentic ASR outputs to expand scenario coverage; (3) Multi-agent adversarial synthesis that simulates emerging fraud tactics through predefined communication scenarios and fraud typologies. The generated dataset contains 28,511 rigorously processed speech-text pairs, complete with detailed annotations for fraud reasoning. The dataset is divided into three tasks: scenario classification, fraud detection, fraud type classification. Furthermore, we construct TeleAntiFraud-Bench, a standardized evaluation benchmark comprising proportionally sampled instances from the dataset, to facilitate systematic testing of model performance on telecom fraud detection tasks. We also contribute a production-optimized supervised fine-tuning (SFT) model trained on hybrid real/synthetic data, while open-sourcing the data processing framework to enable community-driven dataset expansion. This work establishes a foundational framework for multimodal anti-fraud research while addressing critical challenges in data privacy and scenario diversity. The project will be released at https://github.com/JimmyMa99/TeleAntiFraud.

♻ ☆ NTSEBENCH: Cognitive Reasoning Benchmark for Vision Language Models

Pranshu Pandya, Vatsal Gupta, Agney S Talwarr, Tushar Kataria, Dan Roth, Vivek Gupta

Cognitive textual and visual reasoning tasks, including puzzles, series, and analogies, demand the ability to quickly reason, decipher, and evaluate patterns both textually and spatially. Due to extensive training on vast amounts of human-curated data, LLMs and VLMs excel in common-sense reasoning tasks, however still struggle with more complex reasoning that demands deeper cognitive understanding. We introduce NTSEBench, a new dataset designed to evaluate cognitive multi-modal reasoning and problem-solving skills of large models. The dataset contains 2728 multiple-choice questions, accompanied by a total of 4,642 images, categorized into 26 different types. These questions are drawn from the nationwide NTSE examination in India and feature a mix of visual and textual general aptitude challenges, designed to assess intelligence and critical thinking skills beyond mere rote learning. We establish baselines on the dataset using state-of-the-art LLMs and VLMs. To facilitate a comparison between open source and propriety models, we propose four distinct modeling strategies to handle different modalities -- text and images -- in the dataset instances.

comment: 28 pages, 3 figures, 12 tables

♻ ☆ STORYSUMM: Evaluating Faithfulness in Story Summarization EMNLP

Melanie Subbiah, Faisal Ladhak, Akankshya Mishra, Griffin Adams, Lydia B. Chilton, Kathleen McKeown

Human evaluation has been the gold standard for checking faithfulness in abstractive summarization. However, with a challenging source domain like narrative, multiple annotators can agree a summary is faithful, while missing details that are obvious errors only once pointed out. We therefore introduce a new dataset, STORYSUMM, comprising LLM summaries of short stories with localized faithfulness labels and error explanations. This benchmark is for evaluation methods, testing whether a given method can detect challenging inconsistencies. Using this dataset, we first show that any one human annotation protocol is likely to miss inconsistencies, and we advocate for pursuing a range of methods when establishing ground truth for a summarization dataset. We finally test recent automatic metrics and find that none of them achieve more than 70% balanced accuracy on this task, demonstrating that it is a challenging benchmark for future work in faithfulness evaluation.

comment: EMNLP Main 2024

♻ ☆ ASP-based Multi-shot Reasoning via DLV2 with Incremental Grounding

Francesco Calimeri, Giovambattista Ianni, Francesco Pacenza, Simona Perri, Jessica Zangari

DLV2 is an AI tool for Knowledge Representation and Reasoning which supports Answer Set Programming (ASP) - a logic-based declarative formalism, successfully used in both academic and industrial applications. Given a logic program modelling a computational problem, an execution of DLV2 produces the so-called answer sets that correspond one-to-one to the solutions to the problem at hand. The computational process of DLV2 relies on the typical Ground & Solve approach where the grounding step transforms the input program into a new, equivalent ground program, and the subsequent solving step applies propositional algorithms to search for the answer sets. Recently, emerging applications in contexts such as stream reasoning and event processing created a demand for multi-shot reasoning: here, the system is expected to be reactive while repeatedly executed over rapidly changing data. In this work, we present a new incremental reasoner obtained from the evolution of DLV2 towards iterated reasoning. Rather than restarting the computation from scratch, the system remains alive across repeated shots, and it incrementally handles the internal grounding process. At each shot, the system reuses previous computations for building and maintaining a large, more general ground program, from which a smaller yet equivalent portion is determined and used for computing answer sets. Notably, the incremental process is performed in a completely transparent fashion for the user. We describe the system, its usage, its applicability and performance in some practically relevant domains. Under consideration in Theory and Practice of Logic Programming (TPLP).

comment: Under consideration in Theory and Practice of Logic Programming (TPLP)

♻ ☆ A Survey on Unlearnable Data

Jiahao Li, Yiqiang Chen, Yunbing Xing, Yang Gu, Xiangyuan Lan

Unlearnable data (ULD) has emerged as an innovative defense technique to prevent machine learning models from learning meaningful patterns from specific data, thus protecting data privacy and security. By introducing perturbations to the training data, ULD degrades model performance, making it difficult for unauthorized models to extract useful representations. Despite the growing significance of ULD, existing surveys predominantly focus on related fields, such as adversarial attacks and machine unlearning, with little attention given to ULD as an independent area of study. This survey fills that gap by offering a comprehensive review of ULD, examining unlearnable data generation methods, public benchmarks, evaluation metrics, theoretical foundations and practical applications. We compare and contrast different ULD approaches, analyzing their strengths, limitations, and trade-offs related to unlearnability, imperceptibility, efficiency and robustness. Moreover, we discuss key challenges, such as balancing perturbation imperceptibility with model degradation and the computational complexity of ULD generation. Finally, we highlight promising future research directions to advance the effectiveness and applicability of ULD, underscoring its potential to become a crucial tool in the evolving landscape of data protection in machine learning.

comment: 31 pages, 3 figures, Code in https://github.com/LiJiahao-Alex/Awesome-UnLearnable-Data

♻ ☆ LLM-Human Pipeline for Cultural Context Grounding of Conversations NAACL 2025

Rajkumar Pujari, Dan Goldwasser

Conversations often adhere to well-understood social norms that vary across cultures. For example, while "addressing parents by name" is commonplace in the West, it is rare in most Asian cultures. Adherence or violation of such norms often dictates the tenor of conversations. Humans are able to navigate social situations requiring cultural awareness quite adeptly. However, it is a hard task for NLP models. In this paper, we tackle this problem by introducing a "Cultural Context Schema" for conversations. It comprises (1) conversational information such as emotions, dialogue acts, etc., and (2) cultural information such as social norms, violations, etc. We generate ~110k social norm and violation descriptions for ~23k conversations from Chinese culture using LLMs. We refine them using automated verification strategies which are evaluated against culturally aware human judgements. We organize these descriptions into meaningful structures we call "Norm Concepts", using an interactive human-in-loop framework. We ground the norm concepts and the descriptions in conversations using symbolic annotation. Finally, we use the obtained dataset for downstream tasks such as emotion, sentiment, and dialogue act detection. We show that it significantly improves the empirical performance.

comment: Oral at NAACL 2025 Main conference. Albuquerque, USA. Apr 29 - May 4, 2025. 19 pages, 9 figures, 7 tables

♻ ☆ NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals

Jaden Fiotto-Kaufman, Alexander R. Loftus, Eric Todd, Jannik Brinkmann, Koyena Pal, Dmitrii Troitskii, Michael Ripa, Adam Belfki, Can Rager, Caden Juang, Aaron Mueller, Samuel Marks, Arnab Sen Sharma, Francesca Lucchetti, Nikhil Prakash, Carla Brodley, Arjun Guha, Jonathan Bell, Byron C. Wallace, David Bau

We introduce NNsight and NDIF, technologies that work in tandem to enable scientific study of the representations and computations learned by very large neural networks. NNsight is an open-source system that extends PyTorch to introduce deferred remote execution. The National Deep Inference Fabric (NDIF) is a scalable inference service that executes NNsight requests, allowing users to share GPU resources and pretrained models. These technologies are enabled by the Intervention Graph, an architecture developed to decouple experimental design from model runtime. Together, this framework provides transparent and efficient access to the internals of deep neural networks such as very large language models (LLMs) without imposing the cost or complexity of hosting customized models individually. We conduct a quantitative survey of the machine learning literature that reveals a growing gap in the study of the internals of large-scale AI. We demonstrate the design and use of our framework to address this gap by enabling a range of research methods on huge models. Finally, we conduct benchmarks to compare performance with previous approaches. Code, documentation, and tutorials are available at https://nnsight.net/.

comment: Code at https://nnsight.net

♻ ☆ Large Language Models are In-Context Molecule Learners

Jiatong Li, Wei Liu, Zhihao Ding, Wenqi Fan, Yuqiang Li, Qing Li

Large Language Models (LLMs) have demonstrated exceptional performance in biochemical tasks, especially the molecule caption translation task, which aims to bridge the gap between molecules and natural language texts. However, previous methods in adapting LLMs to the molecule-caption translation task required extra domain-specific pre-training stages, suffered weak alignment between molecular and textual spaces, or imposed stringent demands on the scale of LLMs. To resolve the challenges, we propose In-Context Molecule Adaptation (ICMA), as a new paradigm allowing LLMs to learn the molecule-text alignment from context examples via In-Context Molecule Tuning. Specifically, ICMA incorporates the following three stages: Hybrid Context Retrieval, Post-retrieval Re-ranking, and In-context Molecule Tuning. Initially, Hybrid Context Retrieval utilizes BM25 Caption Retrieval and Molecule Graph Retrieval to retrieve similar informative context examples. Additionally, Post-retrieval Re-ranking is composed of Sequence Reversal and Random Walk selection to further improve the quality of retrieval results. Finally, In-Context Molecule Tuning unlocks the in-context learning and reasoning capability of LLMs with the retrieved examples and adapts the parameters of LLMs for better alignment between molecules and texts. Experimental results demonstrate that ICMA can empower LLMs to achieve state-of-the-art or comparable performance without extra training corpora and intricate structures, showing that LLMs are inherently in-context molecule learners.

comment: Accepted by IEEE TKDE

♻ ☆ An Optimistic-Robust Approach for Dynamic Positioning of Omnichannel Inventories

Pavithra Harsha, Shivaram Subramanian, Ali Koc, Mahesh Ramakrishna, Brian Quanz, Dhruv Shah, Chandra Narayanaswami

We introduce a new class of data-driven and distribution-free optimistic-robust bimodal inventory optimization (BIO) strategy to effectively allocate inventory across a retail chain to meet time-varying, uncertain omnichannel demand. The bimodal nature of BIO stems from its ability to balance downside risk, as in traditional Robust Optimization (RO), which focuses on worst-case adversarial demand, with upside potential to enhance average-case performance. This enables BIO to remain as resilient as RO while capturing benefits that would otherwise be lost due to endogenous outliers. Omnichannel inventory planning provides a suitable problem setting for analyzing the effectiveness of BIO's bimodal strategy in managing the tradeoff between lost sales at stores and cross-channel e-commerce fulfillment costs, factors that are inherently asymmetric due to channel-specific behaviors. We provide structural insights about the BIO solution and how it can be tuned to achieve a preferred tradeoff between robustness and the average-case performance. Using a real-world dataset from a large American omnichannel retail chain, a business value assessment during a peak period indicates that BIO outperforms pure RO by 27% in terms of realized average profitability and surpasses other competitive baselines under imperfect distributional information by over 10%. This demonstrates that BIO provides a novel, data-driven, and distribution-free alternative to traditional RO that achieves strong average performance while carefully balancing robustness.

♻ ☆ AI-Powered Bayesian Inference

Veronika Ročková, Sean O'Hagan

The advent of Generative Artificial Intelligence (GAI) has heralded an inflection point that changed how society thinks about knowledge acquisition. While GAI cannot be fully trusted for decision-making, it may still provide valuable information that can be integrated into a decision pipeline. Rather than seeing the lack of certitude and inherent randomness of GAI as a problem, we view it as an opportunity. Indeed, variable answers to given prompts can be leveraged to construct a prior distribution which reflects assuredness of AI predictions. This prior distribution may be combined with tailored datasets for a fully Bayesian analysis with an AI-driven prior. In this paper, we explore such a possibility within a non-parametric Bayesian framework. The basic idea consists of assigning a Dirichlet process prior distribution on the data-generating distribution with AI generative model as its baseline. Hyper-parameters of the prior can be tuned out-of-sample to assess the informativeness of the AI prior. Posterior simulation is achieved by computing a suitably randomized functional on an augmented data that consists of observed (labeled) data as well as fake data whose labels have been imputed using AI. This strategy can be parallelized and rapidly produces iid samples from the posterior by optimization as opposed to sampling from conditionals. Our method enables (predictive) inference and uncertainty quantification leveraging AI predictions in a coherent probabilistic manner.

comment: 37 pages, 4 figures; added additional experiments, asymptotic theory and exposition, corrected typos

♻ ☆ Explainable Bayesian Optimization

Tanmay Chakraborty, Christian Wirth, Christin Seifert

Manual parameter tuning of cyber-physical systems is a common practice, but it is labor-intensive. Bayesian Optimization (BO) offers an automated alternative, yet its black-box nature reduces trust and limits human-BO collaborative system tuning. Experts struggle to interpret BO recommendations due to the lack of explanations. This paper addresses the post-hoc BO explainability problem for cyber-physical systems. We introduce TNTRules (Tune-No-Tune Rules), a novel algorithm that provides both global and local explanations for BO recommendations. TNTRules generates actionable rules and visual graphs, identifying optimal solution bounds and ranges, as well as potential alternative solutions. Unlike existing explainable AI (XAI) methods, TNTRules is tailored specifically for BO, by encoding uncertainty via a variance pruning technique and hierarchical agglomerative clustering. A multi-objective optimization approach allows maximizing explanation quality. We evaluate TNTRules using established XAI metrics (Correctness, Completeness, and Compactness) and compare it against adapted baseline methods. The results demonstrate that TNTRules generates high-fidelity, compact, and complete explanations, significantly outperforming three baselines on 5 multi-objective testing functions and 2 hyperparameter tuning problems.

♻ ☆ BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games ICLR 2025

Davide Paglieri, Bartłomiej Cupiał, Samuel Coward, Ulyana Piterbarg, Maciej Wolczyk, Akbir Khan, Eduardo Pignatelli, Łukasz Kuciński, Lerrel Pinto, Rob Fergus, Jakob Nicolaus Foerster, Jack Parker-Holder, Tim Rocktäschel

Large Language Models (LLMs) and Vision Language Models (VLMs) possess extensive knowledge and exhibit promising reasoning abilities, however, they still struggle to perform well in complex, dynamic environments. Real-world tasks require handling intricate interactions, advanced spatial reasoning, long-term planning, and continuous exploration of new strategies-areas in which we lack effective methodologies for comprehensively evaluating these capabilities. To address this gap, we introduce BALROG, a novel benchmark designed to assess the agentic capabilities of LLMs and VLMs through a diverse set of challenging games. Our benchmark incorporates a range of existing reinforcement learning environments with varying levels of difficulty, including tasks that are solvable by non-expert humans in seconds to extremely challenging ones that may take years to master (e.g., the NetHack Learning Environment). We devise fine-grained metrics to measure performance and conduct an extensive evaluation of several popular open-source and closed-source LLMs and VLMs. Our findings indicate that while current models achieve partial success in the easier games, they struggle significantly with more challenging tasks. Notably, we observe severe deficiencies in vision-based decision-making, as several models perform worse when visual representations of the environments are provided. We release BALROG as an open and user-friendly benchmark to facilitate future research and development in the agentic community. Code and Leaderboard at balrogai.com.

comment: Published as a conference paper at ICLR 2025

♻ ☆ DELTA: Decomposed Efficient Long-Term Robot Task Planning using Large Language Models ICRA 2025

Yuchen Liu, Luigi Palmieri, Sebastian Koch, Ilche Georgievski, Marco Aiello

Recent advancements in Large Language Models (LLMs) have sparked a revolution across many research fields. In robotics, the integration of common-sense knowledge from LLMs into task and motion planning has drastically advanced the field by unlocking unprecedented levels of context awareness. Despite their vast collection of knowledge, large language models may generate infeasible plans due to hallucinations or missing domain information. To address these challenges and improve plan feasibility and computational efficiency, we introduce DELTA, a novel LLM-informed task planning approach. By using scene graphs as environment representations within LLMs, DELTA achieves rapid generation of precise planning problem descriptions. To enhance planning performance, DELTA decomposes long-term task goals with LLMs into an autoregressive sequence of sub-goals, enabling automated task planners to efficiently solve complex problems. In our extensive evaluation, we show that DELTA enables an efficient and fully automatic task planning pipeline, achieving higher planning success rates and significantly shorter planning times compared to the state of the art. Project webpage: https://delta-llm.github.io/

comment: Accepted at ICRA 2025

♻ ☆ Knowledge-Aware Iterative Retrieval for Multi-Agent Systems

Seyoung Song

We introduce a novel large language model (LLM)-driven agent framework, which iteratively refines queries and filters contextual evidence by leveraging dynamically evolving knowledge. A defining feature of the system is its decoupling of external sources from an internal knowledge cache that is progressively updated to guide both query generation and evidence selection. This design mitigates bias-reinforcement loops and enables dynamic, trackable search exploration paths, thereby optimizing the trade-off between exploring diverse information and maintaining accuracy through autonomous agent decision-making. Our approach is evaluated on a broad range of open-domain question answering benchmarks, including multi-step tasks that mirror real-world scenarios where integrating information from multiple sources is critical, especially given the vulnerabilities of LLMs that lack explicit reasoning or planning capabilities. The results show that the proposed system not only outperforms single-step baselines regardless of task difficulty but also, compared to conventional iterative retrieval methods, demonstrates pronounced advantages in complex tasks through precise evidence-based reasoning and enhanced efficiency. The proposed system supports both competitive and collaborative sharing of updated context, enabling multi-agent extension. The benefits of multi-agent configurations become especially prominent as task difficulty increases. The number of convergence steps scales with task difficulty, suggesting cost-effective scalability.

♻ ☆ The Computational Complexity of Circuit Discovery for Inner Interpretability ICLR 2025

Federico Adolfi, Martina G. Vilas, Todd Wareham

Many proposed applications of neural networks in machine learning, cognitive/brain science, and society hinge on the feasibility of inner interpretability via circuit discovery. This calls for empirical and theoretical explorations of viable algorithmic options. Despite advances in the design and testing of heuristics, there are concerns about their scalability and faithfulness at a time when we lack understanding of the complexity properties of the problems they are deployed to solve. To address this, we study circuit discovery with classical and parameterized computational complexity theory: (1) we describe a conceptual scaffolding to reason about circuit finding queries in terms of affordances for description, explanation, prediction and control; (2) we formalize a comprehensive set of queries for mechanistic explanation, and propose a formal framework for their analysis; (3) we use it to settle the complexity of many query variants and relaxations of practical interest on multi-layer perceptrons. Our findings reveal a challenging complexity landscape. Many queries are intractable, remain fixed-parameter intractable relative to model/circuit features, and inapproximable under additive, multiplicative, and probabilistic approximation schemes. To navigate this landscape, we prove there exist transformations to tackle some of these hard problems with better-understood heuristics, and prove the tractability or fixed-parameter tractability of more modest queries which retain useful affordances. This framework allows us to understand the scope and limits of interpretability queries, explore viable options, and compare their resource demands on existing and future architectures.

comment: ICLR 2025 (Spotlight)

♻ ☆ TOBUGraph: Knowledge Graph-Based Retrieval for Enhanced LLM Performance Beyond RAG

Savini Kashmira, Jayanaka L. Dantanarayana, Joshua Brodsky, Ashish Mahendra, Yiping Kang, Krisztian Flautner, Lingjia Tang, Jason Mars

Retrieval-Augmented Generation (RAG) is one of the leading and most widely used techniques for enhancing LLM retrieval capabilities, but it still faces significant limitations in commercial use cases. RAG primarily relies on the query-chunk text-to-text similarity in the embedding space for retrieval and can fail to capture deeper semantic relationships across chunks, is highly sensitive to chunking strategies, and is prone to hallucinations. To address these challenges, we propose TOBUGraph, a graph-based retrieval framework that first constructs the knowledge graph from unstructured data dynamically and automatically. Using LLMs, TOBUGraph extracts structured knowledge and diverse relationships among data, going beyond RAG's text-to-text similarity. Retrieval is achieved through graph traversal, leveraging the extracted relationships and structures to enhance retrieval accuracy, eliminating the need for chunking configurations while reducing hallucination. We demonstrate TOBUGraph's effectiveness in TOBU, a real-world application in production for personal memory organization and retrieval. Our evaluation using real user data demonstrates that TOBUGraph outperforms multiple RAG implementations in both precision and recall, significantly improving user experience through improved retrieval accuracy.

♻ ☆ Enhanced Controllability of Diffusion Models via Feature Disentanglement and Realism-Enhanced Sampling Methods ECCV 2024

Wonwoong Cho, Hareesh Ravi, Midhun Harikumar, Vinh Khuc, Krishna Kumar Singh, Jingwan Lu, David I. Inouye, Ajinkya Kale

As Diffusion Models have shown promising performance, a lot of efforts have been made to improve the controllability of Diffusion Models. However, how to train Diffusion Models to have the disentangled latent spaces and how to naturally incorporate the disentangled conditions during the sampling process have been underexplored. In this paper, we present a training framework for feature disentanglement of Diffusion Models (FDiff). We further propose two sampling methods that can boost the realism of our Diffusion Models and also enhance the controllability. Concisely, we train Diffusion Models conditioned on two latent features, a spatial content mask, and a flattened style embedding. We rely on the inductive bias of the denoising process of Diffusion Models to encode pose/layout information in the content feature and semantic/style information in the style feature. Regarding the sampling methods, we first generalize Composable Diffusion Models (GCDM) by breaking the conditional independence assumption to allow for some dependence between conditional inputs, which is shown to be effective in realistic generation in our experiments. Second, we propose timestep-dependent weight scheduling for content and style features to further improve the performance. We also observe better controllability of our proposed methods compared to existing methods in image manipulation and image translation.

comment: ECCV 2024; Code will be opened after a patent application is granted

♻ ☆ Att-Adapter: A Robust and Precise Domain-Specific Multi-Attributes T2I Diffusion Adapter via Conditional Variational Autoencoder

Wonwoong Cho, Yan-Ying Chen, Matthew Klenk, David I. Inouye, Yanxia Zhang

Text-to-Image (T2I) Diffusion Models have achieved remarkable performance in generating high quality images. However, enabling precise control of continuous attributes, especially multiple attributes simultaneously, in a new domain (e.g., numeric values like eye openness or car width) with text-only guidance remains a significant challenge. To address this, we introduce the Attribute (Att) Adapter, a novel plug-and-play module designed to enable fine-grained, multi-attributes control in pretrained diffusion models. Our approach learns a single control adapter from a set of sample images that can be unpaired and contain multiple visual attributes. The Att-Adapter leverages the decoupled cross attention module to naturally harmonize the multiple domain attributes with text conditioning. We further introduce Conditional Variational Autoencoder (CVAE) to the Att-Adapter to mitigate overfitting, matching the diverse nature of the visual world. Evaluations on two public datasets show that Att-Adapter outperforms all LoRA-based baselines in controlling continuous attributes. Additionally, our method enables a broader control range and also improves disentanglement across multiple attributes, surpassing StyleGAN-based techniques. Notably, Att-Adapter is flexible, requiring no paired synthetic data for training, and is easily scalable to multiple attributes within a single model.

♻ ☆ Statistically Testing Training Data for Unwanted Error Patterns using Rule-Oriented Regression

Stefan Rass, Martin Dallinger

Artificial intelligence models trained from data can only be as good as the underlying data is. Biases in training data propagating through to the output of a machine learning model are a well-documented and well-understood phenomenon, but the machinery to prevent these undesired effects is much less developed. Efforts to ensure data is clean during collection, such as using bias-aware sampling, are most effective when the entity controlling data collection also trains the AI. In cases where the data is already available, how do we find out if the data was already manipulated, i.e., ``poisoned'', so that an undesired behavior would be trained into a machine learning model? This is a challenge fundamentally different to (just) improving approximation accuracy or efficiency, and we provide a method to test training data for flaws, to establish a trustworthy ground-truth for a subsequent training of machine learning models (of any kind). Unlike the well-studied problem of approximating data using fuzzy rules that are generated from the data, our method hinges on a prior definition of rules to happen before seeing the data to be tested. Therefore, the proposed method can also discover hidden error patterns, which may also have substantial influence. Our approach extends the abilities of conventional statistical testing by letting the ``test-condition'' be any Boolean condition to describe a pattern in the data, whose presence we wish to determine. The method puts fuzzy inference into a regression model, to get the best of the two: explainability from fuzzy logic with statistical properties and diagnostics from the regression, and finally also being applicable to ``small data'', hence not requiring large datasets as deep learning methods do. We provide an open source implementation for demonstration and experiments.

♻ ☆ Class-Dependent Perturbation Effects in Evaluating Time Series Attributions

Gregor Baer, Isel Grau, Chao Zhang, Pieter Van Gorp

As machine learning models become increasingly prevalent in time series applications, Explainable Artificial Intelligence (XAI) methods are essential for understanding their predictions. Within XAI, feature attribution methods aim to identify which input features contribute the most to a model's prediction, with their evaluation typically relying on perturbation-based metrics. Through systematic empirical analysis across multiple datasets, model architectures, and perturbation strategies, we reveal previously overlooked class-dependent effects in these metrics: they show varying effectiveness across classes, achieving strong results for some while remaining less sensitive to others. In particular, we find that the most effective perturbation strategies often demonstrate the most pronounced class differences. Our analysis suggests that these effects arise from the learned biases of classifiers, indicating that perturbation-based evaluation may reflect specific model behaviors rather than intrinsic attribution quality. We propose an evaluation framework with a class-aware penalty term to help assess and account for these effects in evaluating feature attributions, offering particular value for class-imbalanced datasets. Although our analysis focuses on time series classification, these class-dependent effects likely extend to other structured data domains where perturbation-based evaluation is common.

comment: Accepted at The World Conference on eXplainable Artificial Intelligence (XAI-2025)

♻ ☆ Scalable Safe Multi-Agent Reinforcement Learning for Multi-Agent System

Haikuo Du, Fandi Gou, Yunze Cai

Safety and scalability are two critical challenges faced by practical Multi-Agent Systems (MAS). However, existing Multi-Agent Reinforcement Learning (MARL) algorithms that rely solely on reward shaping are ineffective in ensuring safety, and their scalability is rather limited due to the fixed-size network output. To address these issues, we propose a novel framework, Scalable Safe MARL (SS-MARL), to enhance the safety and scalability of MARL methods. Leveraging the inherent graph structure of MAS, we design a multi-layer message passing network to aggregate local observations and communications of varying sizes. Furthermore, we develop a constrained joint policy optimization method in the setting of local observation to improve safety. Simulation experiments demonstrate that SS-MARL achieves a better trade-off between optimality and safety compared to baselines, and its scalability significantly outperforms the latest methods in scenarios with a large number of agents.

♻ ☆ Reasoning-SQL: Reinforcement Learning with SQL Tailored Partial Rewards for Reasoning-Enhanced Text-to-SQL

Mohammadreza Pourreza, Shayan Talaei, Ruoxi Sun, Xingchen Wan, Hailong Li, Azalia Mirhoseini, Amin Saberi, Sercan "O. Arik

Text-to-SQL is a challenging task involving multiple reasoning-intensive subtasks, including natural language understanding, database schema comprehension, and precise SQL query formulation. Existing approaches often rely on handcrafted reasoning paths with inductive biases that can limit their overall effectiveness. Motivated by the recent success of reasoning-enhanced models such as DeepSeek R1 and OpenAI o1, which effectively leverage reward-driven self-exploration to enhance reasoning capabilities and generalization, we propose a novel set of partial rewards tailored specifically for the Text-to-SQL task. Our reward set includes schema-linking, AI feedback, n-gram similarity, and syntax check, explicitly designed to address the reward sparsity issue prevalent in reinforcement learning (RL). Leveraging group relative policy optimization (GRPO), our approach explicitly encourages large language models (LLMs) to develop intrinsic reasoning skills necessary for accurate SQL query generation. With models of different sizes, we demonstrate that RL-only training with our proposed rewards consistently achieves higher accuracy and superior generalization compared to supervised fine-tuning (SFT). Remarkably, our RL-trained 14B-parameter model significantly outperforms larger proprietary models, e.g. o3-mini by 4% and Gemini-1.5-Pro-002 by 3% on the BIRD benchmark. These highlight the efficacy of our proposed RL-training framework with partial rewards for enhancing both accuracy and reasoning capabilities in Text-to-SQL tasks.

comment: Mohammadreza Pourreza and Shayan Talaei contributed equally to this work

♻ ☆ Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning

Zhenni Bi, Kai Han, Chuanjian Liu, Yehui Tang, Yunhe Wang

Large Language Models (LLMs) have demonstrated remarkable abilities across various language tasks, but solving complex reasoning problems remains a significant challenge. While existing methods, such as Chain-of-Thought (CoT) and Tree-of-Thought (ToT), enhance reasoning by decomposing problems or structuring prompts, they typically perform a single pass of reasoning and may fail to revisit flawed paths, compromising accuracy. To address this limitation, we propose a novel reasoning framework called Forest-of-Thought (FoT), which integrates multiple reasoning trees to leverage collective decision-making for solving complex logical problems. FoT employs sparse activation strategies to select the most relevant reasoning paths, improving both efficiency and accuracy. Additionally, we introduce a dynamic self-correction strategy that enables real-time error correction, along with consensus-guided decision-making strategies to optimize both correctness and computational resources. Experimental results demonstrate that the FoT framework, combined with these strategies, significantly enhances the reasoning capabilities of LLMs, enabling them to solve complex tasks with greater precision and efficiency. Code will be available at https://github.com/iamhankai/Forest-of-Thought.

comment: Preprint

♻ ☆ PICLe: Pseudo-Annotations for In-Context Learning in Low-Resource Named Entity Detection NAACL2025

Sepideh Mamooler, Syrielle Montariol, Alexander Mathis, Antoine Bosselut

In-context learning (ICL) enables Large Language Models (LLMs) to perform tasks using few demonstrations, facilitating task adaptation when labeled examples are hard to obtain. However, ICL is sensitive to the choice of demonstrations, and it remains unclear which demonstration attributes enable in-context generalization. In this work, we conduct a perturbation study of in-context demonstrations for low-resource Named Entity Detection (NED). Our surprising finding is that in-context demonstrations with partially correct annotated entity mentions can be as effective for task transfer as fully correct demonstrations. Based off our findings, we propose Pseudo-annotated In-Context Learning (PICLe), a framework for in-context learning with noisy, pseudo-annotated demonstrations. PICLe leverages LLMs to annotate many demonstrations in a zero-shot first pass. We then cluster these synthetic demonstrations, sample specific sets of in-context demonstrations from each cluster, and predict entity mentions using each set independently. Finally, we use self-verification to select the final set of entity mentions. We evaluate PICLe on five biomedical NED datasets and show that, with zero human annotation, PICLe outperforms ICL in low-resource settings where limited gold examples can be used as in-context demonstrations.

comment: In Proceedings of NAACL2025

♻ ☆ A Graph-to-Text Approach to Knowledge-Grounded Response Generation in Human-Robot Interaction

Nicholas Thomas Walker, Stefan Ultes, Pierre Lison

Knowledge graphs are often used to represent structured information in a flexible and efficient manner, but their use in situated dialogue remains under-explored. This paper presents a novel conversational model for human--robot interaction that rests upon a graph-based representation of the dialogue state. The knowledge graph representing the dialogue state is continuously updated with new observations from the robot sensors, including linguistic, situated and multimodal inputs, and is further enriched by other modules, in particular for spatial understanding. The neural conversational model employed to respond to user utterances relies on a simple but effective graph-to-text mechanism that traverses the dialogue state graph and converts the traversals into a natural language form. This conversion of the state graph into text is performed using a set of parameterized functions, and the values for those parameters are optimized based on a small set of Wizard-of-Oz interactions. After this conversion, the text representation of the dialogue state graph is included as part of the prompt of a large language model used to decode the agent response. The proposed approach is empirically evaluated through a user study with a humanoid robot that acts as conversation partner to evaluate the impact of the graph-to-text mechanism on the response generation. After moving a robot along a tour of an indoor environment, participants interacted with the robot using spoken dialogue and evaluated how well the robot was able to answer questions about what the robot observed during the tour. User scores show a statistically significant improvement in the perceived factuality of the robot responses when the graph-to-text approach is employed, compared to a baseline using inputs structured as semantic triples.

comment: Submitted to Dialogue & Discourse 2023

♻ ☆ HRET: A Self-Evolving LLM Evaluation Toolkit for Korean

Hanwool Lee, Soo Yong Kim, Dasol Choi, SangWon Baek, Seunghyeok Hong, Ilgyun Jeong, Inseon Hwang, Naeun Lee, Guijin Son

Recent advancements in Korean large language models (LLMs) have spurred numerous benchmarks and evaluation methodologies, yet the lack of a standardized evaluation framework has led to inconsistent results and limited comparability. To address this, we introduce HRET Haerae Evaluation Toolkit, an open-source, self-evolving evaluation framework tailored specifically for Korean LLMs. HRET unifies diverse evaluation methods, including logit-based scoring, exact-match, language-inconsistency penalization, and LLM-as-a-Judge assessments. Its modular, registry-based architecture integrates major benchmarks (HAE-RAE Bench, KMMLU, KUDGE, HRM8K) and multiple inference backends (vLLM, HuggingFace, OpenAI-compatible endpoints). With automated pipelines for continuous evolution, HRET provides a robust foundation for reproducible, fair, and transparent Korean NLP research.

♻ ☆ QualiSpeech: A Speech Quality Assessment Dataset with Natural Language Reasoning and Descriptions

Siyin Wang, Wenyi Yu, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Lu Lu, Yu Tsao, Junichi Yamagishi, Yuxuan Wang, Chao Zhang

This paper explores a novel perspective to speech quality assessment by leveraging natural language descriptions, offering richer, more nuanced insights than traditional numerical scoring methods. Natural language feedback provides instructive recommendations and detailed evaluations, yet existing datasets lack the comprehensive annotations needed for this approach. To bridge this gap, we introduce QualiSpeech, a comprehensive low-level speech quality assessment dataset encompassing 11 key aspects and detailed natural language comments that include reasoning and contextual insights. Additionally, we propose the QualiSpeech Benchmark to evaluate the low-level speech understanding capabilities of auditory large language models (LLMs). Experimental results demonstrate that finetuned auditory LLMs can reliably generate detailed descriptions of noise and distortion, effectively identifying their types and temporal characteristics. The results further highlight the potential for incorporating reasoning to enhance the accuracy and reliability of quality assessments. The dataset will be released at https://huggingface.co/datasets/tsinghua-ee/QualiSpeech.

comment: 23 pages, 16 figures

♻ ☆ Sabiá-3 Technical Report

Hugo Abonizio, Thales Sales Almeida, Thiago Laitz, Roseval Malaquias Junior, Giovana Kerche Bonás, Rodrigo Nogueira, Ramon Pires

This report presents Sabi\'a-3, our new flagship language model, and Sabiazinho-3, a more cost-effective sibling. The models were trained on a large brazilian-centric corpus. Evaluations across diverse professional and academic benchmarks show a strong performance on Portuguese and Brazil-related tasks. Sabi\'a-3 shows large improvements in comparison to our previous best of model, Sabia-2 Medium, especially in reasoning-intensive tasks. Notably, Sabi\'a-3's average performance matches frontier LLMs, while it is offered at a three to four times lower cost per token, reinforcing the benefits of domain specialization.

♻ ☆ Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation

Han Xue, Jieji Ren, Wendi Chen, Gu Zhang, Yuan Fang, Guoying Gu, Huazhe Xu, Cewu Lu

Humans can accomplish complex contact-rich tasks using vision and touch, with highly reactive capabilities such as quick adjustments to environmental changes and adaptive control of contact forces; however, this remains challenging for robots. Existing visual imitation learning (IL) approaches rely on action chunking to model complex behaviors, which lacks the ability to respond instantly to real-time tactile feedback during the chunk execution. Furthermore, most teleoperation systems struggle to provide fine-grained tactile / force feedback, which limits the range of tasks that can be performed. To address these challenges, we introduce TactAR, a low-cost teleoperation system that provides real-time tactile feedback through Augmented Reality (AR), along with Reactive Diffusion Policy (RDP), a novel slow-fast visual-tactile imitation learning algorithm for learning contact-rich manipulation skills. RDP employs a two-level hierarchy: (1) a slow latent diffusion policy for predicting high-level action chunks in latent space at low frequency, (2) a fast asymmetric tokenizer for closed-loop tactile feedback control at high frequency. This design enables both complex trajectory modeling and quick reactive behavior within a unified framework. Through extensive evaluation across three challenging contact-rich tasks, RDP significantly improves performance compared to state-of-the-art visual IL baselines through rapid response to tactile / force feedback. Furthermore, experiments show that RDP is applicable across different tactile / force sensors. Code and videos are available on https://reactive-diffusion-policy.github.io.

♻ ☆ Nonhuman Primate Brain Tissue Segmentation Using a Transfer Learning Approach

Zhen Lin, Hongyu Yuan, Richard Barcus, Qing Lyu, Sucheta Chakravarty, Megan E. Lipford, Carol A. Shively, Suzanne Craft, Mohammad Kawas, Jeongchul Kim, Christopher T. Whitlow

Non-human primates (NHPs) serve as critical models for understanding human brain function and neurological disorders due to their close evolutionary relationship with humans. Accurate brain tissue segmentation in NHPs is critical for understanding neurological disorders, but challenging due to the scarcity of annotated NHP brain MRI datasets, the small size of the NHP brain, the limited resolution of available imaging data and the anatomical differences between human and NHP brains. To address these challenges, we propose a novel approach utilizing STU-Net with transfer learning to leverage knowledge transferred from human brain MRI data to enhance segmentation accuracy in the NHP brain MRI, particularly when training data is limited. The combination of STU-Net and transfer learning effectively delineates complex tissue boundaries and captures fine anatomical details specific to NHP brains. Notably, our method demonstrated improvement in segmenting small subcortical structures such as putamen and thalamus that are challenging to resolve with limited spatial resolution and tissue contrast, and achieved DSC of over 0.88, IoU over 0.8 and HD95 under 7. This study introduces a robust method for multi-class brain tissue segmentation in NHPs, potentially accelerating research in evolutionary neuroscience and preclinical studies of neurological disorders relevant to human health.

♻ ☆ Machine Unlearning Fails to Remove Data Poisoning Attacks ICLR 2025

Martin Pawelczyk, Jimmy Z. Di, Yiwei Lu, Ayush Sekhari, Gautam Kamath, Seth Neel

We revisit the efficacy of several practical methods for approximate machine unlearning developed for large-scale deep learning. In addition to complying with data deletion requests, one often-cited potential application for unlearning methods is to remove the effects of poisoned data. We experimentally demonstrate that, while existing unlearning methods have been demonstrated to be effective in a number of settings, they fail to remove the effects of data poisoning across a variety of types of poisoning attacks (indiscriminate, targeted, and a newly-introduced Gaussian poisoning attack) and models (image classifiers and LLMs); even when granted a relatively large compute budget. In order to precisely characterize unlearning efficacy, we introduce new evaluation metrics for unlearning based on data poisoning. Our results suggest that a broader perspective, including a wider variety of evaluations, are required to avoid a false sense of confidence in machine unlearning procedures for deep learning without provable guarantees. Moreover, while unlearning methods show some signs of being useful to efficiently remove poisoned data without having to retrain, our work suggests that these methods are not yet ``ready for prime time,'' and currently provide limited benefit over retraining.

comment: Published at ICLR 2025

♻ ☆ Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning

Gabriele Dominici, Pietro Barbiero, Mateo Espinosa Zarlenga, Alberto Termine, Martin Gjoreski, Giuseppe Marra, Marc Langheinrich

Causal opacity denotes the difficulty in understanding the "hidden" causal structure underlying the decisions of deep neural network (DNN) models. This leads to the inability to rely on and verify state-of-the-art DNN-based systems, especially in high-stakes scenarios. For this reason, circumventing causal opacity in DNNs represents a key open challenge at the intersection of deep learning, interpretability, and causality. This work addresses this gap by introducing Causal Concept Graph Models (Causal CGMs), a class of interpretable models whose decision-making process is causally transparent by design. Our experiments show that Causal CGMs can: (i) match the generalisation performance of causally opaque models, (ii) enable human-in-the-loop corrections to mispredicted intermediate reasoning steps, boosting not just downstream accuracy after corrections but also the reliability of the explanations provided for specific instances, and (iii) support the analysis of interventional and counterfactual scenarios, thereby improving the model's causal interpretability and supporting the effective verification of its reliability and fairness.

♻ ☆ Exploiting Mixture-of-Experts Redundancy Unlocks Multimodal Generative Abilities

Raman Dutt, Harleen Hanspal, Guoxuan Xia, Petru-Daniel Tudosiu, Alexander Black, Yongxin Yang, Steven McDonagh, Sarah Parisot

In this work, we undertake the challenge of augmenting the existing generative capabilities of pre-trained text-only large language models (LLMs) with multi-modal generation capability while satisfying two core constraints: C1 preserving the preservation of original language generative capabilities with negligible performance degradation, and C2 adhering to a small parameter budget to learn the new modality, ensuring scalability and efficiency. In contrast to current approaches that add dedicated modules, thereby significantly increasing the parameter count, we propose a method that leverages the underutilized capacity inherent in deep models. Specifically, we exploit the parameter redundancy within Mixture-of-Experts (MoEs) as a source of additional capacity for learning a new modality, enabling better parameter efficiency (C1). Moreover, we preserve the original language generation capabilities by applying low-rank adaptation exclusively to the tokens of the new modality (C2). Furthermore, we introduce a novel parameter initialization scheme based on the Gromov-Wasserstein distance to improve convergence and training stability. Through an extensive analysis of the routing mechanism, we uncover the emergence of modality-specific pathways and decreased redundancy within the experts that can efficiently unlock multi-modal generative capabilities. Overall, our method can be seamlessly applied to a wide range of contemporary LLMs, providing a new pathway for transitioning from uni-modal to multi-modal architectures.

♻ ☆ Leveraging Joint Predictive Embedding and Bayesian Inference in Graph Self Supervised Learning

Srinitish Srinivasan, Omkumar CU

Graph representation learning has emerged as a cornerstone for tasks like node classification and link prediction, yet prevailing self-supervised learning (SSL) methods face challenges such as computational inefficiency, reliance on contrastive objectives, and representation collapse. Existing approaches often depend on feature reconstruction, negative sampling, or complex decoders, which introduce training overhead and hinder generalization. Further, current techniques which address such limitations fail to account for the contribution of node embeddings to a certain prediction in the absence of labeled nodes. To address these limitations, we propose a novel joint embedding predictive framework for graph SSL that eliminates contrastive objectives and negative sampling while preserving semantic and structural information. Additionally, we introduce a semantic-aware objective term that incorporates pseudo-labels derived from Gaussian Mixture Models (GMMs), enhancing node discriminability by evaluating latent feature contributions. Extensive experiments demonstrate that our framework outperforms state-of-the-art graph SSL methods across benchmarks, achieving superior performance without contrastive loss or complex decoders. Key innovations include (1) a non-contrastive, view-invariant joint embedding predictive architecture, (2) Leveraging single context and multiple targets relationship between subgraphs, and (3) GMM-based pseudo-label scoring to capture semantic contributions. This work advances graph SSL by offering a computationally efficient, collapse-resistant paradigm that bridges spatial and semantic graph features for downstream tasks. The code for our paper can be found at https://github.com/Deceptrax123/JPEB-GSSL

comment: Preprint. Under Review

♻ ☆ Decomposition of one-layer neural networks via the infinite sum of reproducing kernel Banach spaces

Seungcheol Shin, Myungjoo Kang

In this paper, we define the sum of RKBSs using the characterization theorem of RKBSs and show that the sum of RKBSs is compatible with the direct sum of feature spaces. Moreover, we decompose the integral RKBS into the sum of $p$-norm RKBSs. Finally, we provide applications for the structural understanding of the integral RKBS class.

comment: 22 pages

♻ ☆ MTL-LoRA: Low-Rank Adaptation for Multi-Task Learning

Yaming Yang, Dilxat Muhtar, Yelong Shen, Yuefeng Zhan, Jianfeng Liu, Yujing Wang, Hao Sun, Denvy Deng, Feng Sun, Qi Zhang, Weizhu Chen, Yunhai Tong

Parameter-efficient fine-tuning (PEFT) has been widely employed for domain adaptation, with LoRA being one of the most prominent methods due to its simplicity and effectiveness. However, in multi-task learning (MTL) scenarios, LoRA tends to obscure the distinction between tasks by projecting sparse high-dimensional features from different tasks into the same dense low-dimensional intrinsic space. This leads to task interference and suboptimal performance for LoRA and its variants. To tackle this challenge, we propose MTL-LoRA, which retains the advantages of low-rank adaptation while significantly enhancing MTL capabilities. MTL-LoRA augments LoRA by incorporating additional task-adaptive parameters that differentiate task-specific information and capture shared knowledge across various tasks within low-dimensional spaces. This approach enables pre-trained models to jointly adapt to different target domains with a limited number of trainable parameters. Comprehensive experimental results, including evaluations on public academic benchmarks for natural language understanding, commonsense reasoning, and image-text understanding, as well as real-world industrial text Ads relevance datasets, demonstrate that MTL-LoRA outperforms LoRA and its various variants with comparable or even fewer learnable parameters in MTL setting.

comment: 12 Pages, 4 Figures

♻ ☆ Multilingual Performance of a Multimodal Artificial Intelligence System on Multisubject Physics Concept Inventories

Gerd Kortemeyer, Marina Babayeva, Giulia Polverini, Ralf Widenhorn, Bor Gregorcic

We investigate the multilingual and multimodal performance of a large language model-based artificial intelligence (AI) system, GPT-4o, using a diverse set of physics concept inventories spanning multiple languages and subject categories. The inventories, sourced from the PhysPort website, cover classical physics topics such as mechanics, electromagnetism, optics, and thermodynamics, as well as relativity, quantum mechanics, astronomy, mathematics, and laboratory skills. Unlike previous text-only studies, we uploaded the inventories as images to reflect what a student would see on paper, thereby assessing the system's multimodal functionality. Our results indicate variation in performance across subjects, with laboratory skills standing out as the weakest. We also observe differences across languages, with English and European languages showing the strongest performance. Notably, the relative difficulty of an inventory item is largely independent of the language of the survey. When comparing AI results to existing literature on student performance, we find that the AI system outperforms average post-instruction undergraduate students in all subject categories except laboratory skills. Furthermore, the AI performs worse on items requiring visual interpretation of images than on those that are purely text-based.

♻ ☆ FastRM: An efficient and automatic explainability framework for multimodal generative models

Gabriela Ben-Melech Stan, Estelle Aflalo, Man Luo, Shachar Rosenman, Tiep Le, Sayak Paul, Shao-Yen Tseng, Vasudev Lal

Large Vision Language Models (LVLMs) have demonstrated remarkable reasoning capabilities over textual and visual inputs. However, these models remain prone to generating misinformation. Identifying and mitigating ungrounded responses is crucial for developing trustworthy AI. Traditional explainability methods such as gradient-based relevancy maps, offer insight into the decision process of models, but are often computationally expensive and unsuitable for real-time output validation. In this work, we introduce FastRM, an efficient method for predicting explainable Relevancy Maps of LVLMs. Furthermore, FastRM provides both quantitative and qualitative assessment of model confidence. Experimental results demonstrate that FastRM achieves a 99.8% reduction in computation time and a 44.4% reduction in memory footprint compared to traditional relevancy map generation. FastRM allows explainable AI to be more practical and scalable, thereby promoting its deployment in real-world applications and enabling users to more effectively evaluate the reliability of model outputs.

♻ ☆ PhyT2V: LLM-Guided Iterative Self-Refinement for Physics-Grounded Text-to-Video Generation

Qiyao Xue, Xiangyu Yin, Boyuan Yang, Wei Gao

Text-to-video (T2V) generation has been recently enabled by transformer-based diffusion models, but current T2V models lack capabilities in adhering to the real-world common knowledge and physical rules, due to their limited understanding of physical realism and deficiency in temporal modeling. Existing solutions are either data-driven or require extra model inputs, but cannot be generalizable to out-of-distribution domains. In this paper, we present PhyT2V, a new data-independent T2V technique that expands the current T2V model's capability of video generation to out-of-distribution domains, by enabling chain-of-thought and step-back reasoning in T2V prompting. Our experiments show that PhyT2V improves existing T2V models' adherence to real-world physical rules by 2.3x, and achieves 35% improvement compared to T2V prompt enhancers. The source codes are available at: https://github.com/pittisl/PhyT2V.

comment: 28 pages

♻ ☆ GameVibe: A Multimodal Affective Game Corpus

Matthew Barthet, Maria Kaselimi, Kosmas Pinitas, Konstantinos Makantasis, Antonios Liapis, Georgios N. Yannakakis

As online video and streaming platforms continue to grow, affective computing research has undergone a shift towards more complex studies involving multiple modalities. However, there is still a lack of readily available datasets with high-quality audiovisual stimuli. In this paper, we present GameVibe, a novel affect corpus which consists of multimodal audiovisual stimuli, including in-game behavioural observations and third-person affect traces for viewer engagement. The corpus consists of videos from a diverse set of publicly available gameplay sessions across 30 games, with particular attention to ensure high-quality stimuli with good audiovisual and gameplay diversity. Furthermore, we present an analysis on the reliability of the annotators in terms of inter-annotator agreement.

comment: 12 pages, 5 figures, 1 table

♻ ☆ When Counterfactual Reasoning Fails: Chaos and Real-World Complexity

Yahya Aalaila, Gerrit Großmann, Sumantrak Mukherjee, Jonas Wahl, Sebastian Vollmer

Counterfactual reasoning, a cornerstone of human cognition and decision-making, is often seen as the 'holy grail' of causal learning, with applications ranging from interpreting machine learning models to promoting algorithmic fairness. While counterfactual reasoning has been extensively studied in contexts where the underlying causal model is well-defined, real-world causal modeling is often hindered by model and parameter uncertainty, observational noise, and chaotic behavior. The reliability of counterfactual analysis in such settings remains largely unexplored. In this work, we investigate the limitations of counterfactual reasoning within the framework of Structural Causal Models. Specifically, we empirically investigate \emph{counterfactual sequence estimation} and highlight cases where it becomes increasingly unreliable. We find that realistic assumptions, such as low degrees of model uncertainty or chaotic dynamics, can result in counterintuitive outcomes, including dramatic deviations between predicted and true counterfactual trajectories. This work urges caution when applying counterfactual reasoning in settings characterized by chaos and uncertainty. Furthermore, it raises the question of whether certain systems may pose fundamental limitations on the ability to answer counterfactual questions about their behavior.

♻ ☆ MambaPEFT: Exploring Parameter-Efficient Fine-Tuning for Mamba ICLR2025

Masakazu Yoshimura, Teruaki Hayashi, Yota Maeda

An ecosystem of Transformer-based models has been established by building large models with extensive data. Parameter-efficient fine-tuning (PEFT) is a crucial technology for deploying these models to downstream tasks with minimal cost while achieving effective performance. Recently, Mamba, a State Space Model (SSM)-based model, has attracted attention as a potential alternative to Transformers. While many large-scale Mamba-based models have been proposed, efficiently adapting pre-trained Mamba-based models to downstream tasks remains unexplored. In this paper, we conduct an exploratory analysis of PEFT methods for Mamba. We investigate the effectiveness of existing PEFT methods for Transformers when applied to Mamba. We also modify these methods to better align with the Mamba architecture. Additionally, we propose new Mamba-specific PEFT methods that leverage the distinctive structure of Mamba. Our experiments indicate that PEFT performs more effectively for Mamba than Transformers. Lastly, we demonstrate how to effectively combine multiple PEFT methods and provide a framework that outperforms previous works. To ensure reproducibility, we will release the code after publication.

comment: Accepted to ICLR2025

♻ ☆ BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions ICLR 2025

Terry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex Gu, Zhoujun Cheng, Jiawei Liu, Qian Liu, Zijian Wang, Binyuan Hui, Niklas Muennighoff, David Lo, Daniel Fried, Xiaoning Du, Harm de Vries, Leandro Von Werra

Task automation has been greatly empowered by the recent advances in Large Language Models (LLMs) via Python code, where the tasks ranging from software engineering development to general-purpose reasoning. While current benchmarks have shown that LLMs can solve tasks using programs like human developers, the majority of their evaluations are limited to short and self-contained algorithmic tasks or standalone function calls. Solving challenging and practical tasks requires the capability of utilizing diverse function calls as tools to efficiently implement functionalities like data analysis and web development. In addition, using multiple tools to solve a task needs compositional reasoning by accurately understanding complex instructions. Fulfilling both of these characteristics can pose a great challenge for LLMs.To assess how well LLMs can solve challenging and practical tasks via programs, we introduce BigCodeBench, a benchmark that challenges LLMs to invoke multiple function calls as tools from 139 libraries and 7 domains for 1,140 fine-grained tasks. To evaluate LLMs rigorously, each task encompasses 5.6 test cases with an average branch coverage of 99%. In addition, we propose a natural-language-oriented variant of BigCodeBench, BigCodeBench-Instruct, that automatically transforms the original docstrings into short instructions only with essential information. Our extensive evaluation of 60 LLMs shows that LLMs are not yet capable of following complex instructions to use function calls precisely, with scores up to 60%, significantly lower than the human performance of 97%. The results underscore the need for further advancements in this area.

comment: Accpeted at ICLR 2025 (Oral), built with love by the BigCode community :)

♻ ☆ A Clustering Method with Graph Maximum Decoding Information IJCNN 2024

Xinrun Xu, Manying Lv, Zhanbiao Lian, Yurong Wu, Jin Yan, Shan Jiang, Zhiming Ding

The clustering method based on graph models has garnered increased attention for its widespread applicability across various knowledge domains. Its adaptability to integrate seamlessly with other relevant applications endows the graph model-based clustering analysis with the ability to robustly extract "natural associations" or "graph structures" within datasets, facilitating the modelling of relationships between data points. Despite its efficacy, the current clustering method utilizing the graph-based model overlooks the uncertainty associated with random walk access between nodes and the embedded structural information in the data. To address this gap, we present a novel Clustering method for Maximizing Decoding Information within graph-based models, named CMDI. CMDI innovatively incorporates two-dimensional structural information theory into the clustering process, consisting of two phases: graph structure extraction and graph vertex partitioning. Within CMDI, graph partitioning is reformulated as an abstract clustering problem, leveraging maximum decoding information to minimize uncertainty associated with random visits to vertices. Empirical evaluations on three real-world datasets demonstrate that CMDI outperforms classical baseline methods, exhibiting a superior decoding information ratio (DI-R). Furthermore, CMDI showcases heightened efficiency, particularly when considering prior knowledge (PK). These findings underscore the effectiveness of CMDI in enhancing decoding information quality and computational efficiency, positioning it as a valuable tool in graph-based clustering analyses.

comment: 9 pages, 9 figures, IJCNN 2024

♻ ☆ Evaluating machine learning models for predicting pesticides toxicity to honey bees

Jakub Adamczyk, Jakub Poziemski, Pawel Siedlecki

Small molecules play a critical role in the biomedical, environmental, and agrochemical domains, each with distinct physicochemical requirements and success criteria. Although biomedical research benefits from extensive datasets and established benchmarks, agrochemical data remain scarce, particularly with respect to species-specific toxicity. This work focuses on ApisTox, the most comprehensive dataset of experimentally validated chemical toxicity to the honey bee (Apis mellifera), an ecologically vital pollinator. We evaluate ApisTox using a diverse suite of machine learning approaches, including molecular fingerprints, graph kernels, and graph neural networks, as well as pretrained models. Comparative analysis with medicinal datasets from the MoleculeNet benchmark reveals that ApisTox represents a distinct chemical space. Performance degradation on non-medicinal datasets, such as ApisTox, demonstrates their limited generalizability of current state-of-the-art algorithms trained solely on biomedical data. Our study highlights the need for more diverse datasets and for targeted model development geared toward the agrochemical domain.

♻ ☆ Improving Complex Reasoning with Dynamic Prompt Corruption: A soft prompt Optimization Approach ICLR 2025

Sinan Fan, Liang Xie, Chen Shen, Ge Teng, Xiaosong Yuan, Xiaofeng Zhang, Chenxi Huang, Wenxiao Wang, Xiaofei He, Jieping Ye

Prompt-tuning (PT) for large language models (LLMs) can facilitate the performance on various conventional NLP tasks with significantly fewer trainable parameters. However, our investigation reveals that PT provides limited improvement and may even degrade the primitive performance of LLMs on complex reasoning tasks. Such a phenomenon suggests that soft prompts can positively impact certain instances while negatively affecting others, particularly during the later phases of reasoning. To address these challenges, We first identify an information accumulation within the soft prompts. Through detailed analysis, we demonstrate that this phenomenon is often accompanied by erroneous information flow patterns in the deeper layers of the model, which ultimately lead to incorrect reasoning outcomes. we propose a novel method called Dynamic Prompt Corruption (DPC) to take better advantage of soft prompts in complex reasoning tasks, which dynamically adjusts the influence of soft prompts based on their impact on the reasoning process. Specifically, DPC consists of two stages: Dynamic Trigger and Dynamic Corruption. First, Dynamic Trigger measures the impact of soft prompts, identifying whether beneficial or detrimental. Then, Dynamic Corruption mitigates the negative effects of soft prompts by selectively masking key tokens that interfere with the reasoning process. We validate the proposed approach through extensive experiments on various LLMs and reasoning tasks, including GSM8K, MATH, and AQuA. Experimental results demonstrate that DPC can consistently enhance the performance of PT, achieving 4%-8% accuracy gains compared to vanilla prompt tuning, highlighting the effectiveness of our approach and its potential to enhance complex reasoning in LLMs.

comment: Accepted by ICLR 2025

♻ ☆ Holistic analysis on the sustainability of Federated Learning across AI product lifecycle

Hongliu Cao

In light of emerging legal requirements and policies focused on privacy protection, there is a growing trend of companies across various industries adopting Federated Learning (FL). This decentralized approach involves multiple clients or silos, collaboratively training a global model under the coordination of a central server while utilizing their private local data. Unlike traditional methods that necessitate data sharing and transmission, Cross-Silo FL allows clients to share model updates rather than raw data, thereby enhancing privacy. Despite its growing adoption, the carbon impact associated with Cross-Silo FL remains poorly understood due to the limited research in this area. This study seeks to bridge this gap by evaluating the sustainability of Cross-Silo FL throughout the entire AI product lifecycle, extending the analysis beyond the model training phase alone. We systematically compare this decentralized method with traditional centralized approaches and present a robust quantitative framework for assessing the costs and CO2 emissions in real-world Cross-Silo FL environments. Our findings indicate that the energy consumption and costs of model training are comparable between Cross-Silo Federated Learning and Centralized Learning. However, the additional data transfer and storage requirements inherent in Centralized Learning can result in significant, often overlooked CO2 emissions. Moreover, we introduce an innovative data and application management system that integrates Cross-Silo FL and analytics, aiming at improving the sustainability and economic efficiency of IT enterprises.

comment: Presented in Sophia Summit 2023

♻ ☆ Video-T1: Test-Time Scaling for Video Generation

Fangfu Liu, Hanyang Wang, Yimo Cai, Kaiyan Zhang, Xiaohang Zhan, Yueqi Duan

With the scale capability of increasing training data, model size, and computational cost, video generation has achieved impressive results in digital creation, enabling users to express creativity across various domains. Recently, researchers in Large Language Models (LLMs) have expanded the scaling to test-time, which can significantly improve LLM performance by using more inference-time computation. Instead of scaling up video foundation models through expensive training costs, we explore the power of Test-Time Scaling (TTS) in video generation, aiming to answer the question: if a video generation model is allowed to use non-trivial amount of inference-time compute, how much can it improve generation quality given a challenging text prompt. In this work, we reinterpret the test-time scaling of video generation as a searching problem to sample better trajectories from Gaussian noise space to the target video distribution. Specifically, we build the search space with test-time verifiers to provide feedback and heuristic algorithms to guide searching process. Given a text prompt, we first explore an intuitive linear search strategy by increasing noise candidates at inference time. As full-step denoising all frames simultaneously requires heavy test-time computation costs, we further design a more efficient TTS method for video generation called Tree-of-Frames (ToF) that adaptively expands and prunes video branches in an autoregressive manner. Extensive experiments on text-conditioned video generation benchmarks demonstrate that increasing test-time compute consistently leads to significant improvements in the quality of videos. Project page: https://liuff19.github.io/Video-T1

comment: Project page: https://liuff19.github.io/Video-T1

♻ ☆ MolGround: A Benchmark for Molecular Grounding

Jiaxin Wu, Ting Zhang, Rubing Chen, Wengyu Zhang, Chen Jason Zhang, Xiaoyong Wei, Li Qing

Current molecular understanding approaches predominantly focus on the descriptive aspect of human perception, providing broad, topic-level insights. However, the referential aspect -- linking molecular concepts to specific structural components -- remains largely unexplored. To address this gap, we propose a molecular grounding benchmark designed to evaluate a model's referential abilities. We align molecular grounding with established conventions in NLP, cheminformatics, and molecular science, showcasing the potential of NLP techniques to advance molecular understanding within the AI for Science movement. Furthermore, we constructed the largest molecular understanding benchmark to date, comprising 79k QA pairs, and developed a multi-agent grounding prototype as proof of concept. This system outperforms existing models, including GPT-4o, and its grounding outputs have been integrated to enhance traditional tasks such as molecular captioning and ATC (Anatomical, Therapeutic, Chemical) classification.

♻ ☆ Vision-Language Models for Acute Tuberculosis Diagnosis: A Multimodal Approach Combining Imaging and Clinical Data

Ananya Ganapthy, Praveen Shastry, Naveen Kumarasami, Anandakumar D, Keerthana R, Mounigasri M, Varshinipriya M, Kishore Prasath Venkatesh, Bargava Subramanian, Kalyan Sivasailam

Background: This study introduces a Vision-Language Model (VLM) leveraging SIGLIP and Gemma-3b architectures for automated acute tuberculosis (TB) screening. By integrating chest X-ray images and clinical notes, the model aims to enhance diagnostic accuracy and efficiency, particularly in resource-limited settings. Methods: The VLM combines visual data from chest X-rays with clinical context to generate detailed, context-aware diagnostic reports. The architecture employs SIGLIP for visual encoding and Gemma-3b for decoding, ensuring effective representation of acute TB-specific pathologies and clinical insights. Results: Key acute TB pathologies, including consolidation, cavities, and nodules, were detected with high precision (97percent) and recall (96percent). The model demonstrated strong spatial localization capabilities and robustness in distinguishing TB-positive cases, making it a reliable tool for acute TB diagnosis. Conclusion: The multimodal capability of the VLM reduces reliance on radiologists, providing a scalable solution for acute TB screening. Future work will focus on improving the detection of subtle pathologies and addressing dataset biases to enhance its generalizability and application in diverse global healthcare settings.

comment: 11 pages, 3 figures

♻ ☆ 1-2-3-Go! Policy Synthesis for Parameterized Markov Decision Processes via Decision-Tree Learning and Generalization

Muqsit Azeem, Debraj Chakraborty, Sudeep Kanav, Jan Kretinsky, Mohammadsadegh Mohagheghi, Stefanie Mohr, Maximilian Weininger

Despite the advances in probabilistic model checking, the scalability of the verification methods remains limited. In particular, the state space often becomes extremely large when instantiating parameterized Markov decision processes (MDPs) even with moderate values. Synthesizing policies for such \emph{huge} MDPs is beyond the reach of available tools. We propose a learning-based approach to obtain a reasonable policy for such huge MDPs. The idea is to generalize optimal policies obtained by model-checking small instances to larger ones using decision-tree learning. Consequently, our method bypasses the need for explicit state-space exploration of large models, providing a practical solution to the state-space explosion problem. We demonstrate the efficacy of our approach by performing extensive experimentation on the relevant models from the quantitative verification benchmark set. The experimental results indicate that our policies perform well, even when the size of the model is orders of magnitude beyond the reach of state-of-the-art analysis tools.

comment: Extended version of the paper accepted at VMCAI 2025

♻ ☆ Feed-Forward Bullet-Time Reconstruction of Dynamic Scenes from Monocular Videos

Hanxue Liang, Jiawei Ren, Ashkan Mirzaei, Antonio Torralba, Ziwei Liu, Igor Gilitschenski, Sanja Fidler, Cengiz Oztireli, Huan Ling, Zan Gojcic, Jiahui Huang

Recent advancements in static feed-forward scene reconstruction have demonstrated significant progress in high-quality novel view synthesis. However, these models often struggle with generalizability across diverse environments and fail to effectively handle dynamic content. We present BTimer (short for BulletTimer), the first motion-aware feed-forward model for real-time reconstruction and novel view synthesis of dynamic scenes. Our approach reconstructs the full scene in a 3D Gaussian Splatting representation at a given target ('bullet') timestamp by aggregating information from all the context frames. Such a formulation allows BTimer to gain scalability and generalization by leveraging both static and dynamic scene datasets. Given a casual monocular dynamic video, BTimer reconstructs a bullet-time scene within 150ms while reaching state-of-the-art performance on both static and dynamic scene datasets, even compared with optimization-based approaches.

comment: Project website: https://research.nvidia.com/labs/toronto-ai/bullet-timer/

♻ ☆ Using Language Models to Decipher the Motivation Behind Human Behaviors

Yutong Xie, Qiaozhu Mei, Walter Yuan, Matthew O. Jackson

AI presents a novel tool for deciphering the motivations behind human behaviors. We show that by varying prompts to a large language model, we can elicit a full range of human behaviors in a variety of different scenarios in terms of classic economic games. Then by analyzing which prompts are needed to elicit which behaviors, we can infer (decipher) the motivations behind the human behaviors. We also show how one can analyze the prompts to reveal relationships between the classic economic games, providing new insight into what different economic scenarios induce people to think about. We also show how this deciphering process can be used to understand differences in the behavioral tendencies of different populations.

♻ ☆ Markov $α$-Potential Games

Xin Guo, Xinyu Li, Chinmay Maheshwari, Shankar Sastry, Manxi Wu

We propose a new framework of Markov $\alpha$-potential games to study Markov games. We show that any Markov game with finite-state and finite-action is a Markov $\alpha$-potential game, and establish the existence of an associated $\alpha$-potential function. Any optimizer of an $\alpha$-potential function is shown to be an $\alpha$-stationary Nash equilibrium. We study two important classes of practically significant Markov games, Markov congestion games and the perturbed Markov team games, via the framework of Markov $\alpha$-potential games, with explicit characterization of an upper bound for $\alpha$ and its relation to game parameters. Additionally, we provide a semi-infinite linear programming based formulation to obtain an upper bound for $\alpha$ for any Markov game. Furthermore, we study two equilibrium approximation algorithms, namely the projected gradient-ascent algorithm and the sequential maximum improvement algorithm, along with their Nash regret analysis, and corroborate the results with numerical experiments.

comment: 33 pages, 5 figures

♻ ☆ Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents

Yu Gu, Kai Zhang, Yuting Ning, Boyuan Zheng, Boyu Gou, Tianci Xue, Cheng Chang, Sanjari Srivastava, Yanan Xie, Peng Qi, Huan Sun, Yu Su

Language agents based on large language models (LLMs) have demonstrated great promise in automating web-based tasks. Recent work has shown that incorporating advanced planning algorithms, e.g., tree search, is advantageous over reactive planning for web agents. However, unlike simulated sandbox environments, real-world environments such as the web are rife with irreversible actions. This undermines the feasibility of backtracking, a cornerstone of (tree) search. Overly relying on test-time search also hurts efficiency. We advocate model-based planning for web agents that employs a world model to simulate and deliberate over the outcome of each candidate action before committing to one. We systematically explore this paradigm by (1) Proposing a model-based planning framework, WebDreamer, which employs LLMs to serve as both world models and value functions; (2) Training specialized LLMs as world models with a scalable data synthesis pipeline. Empirical results demonstrate that WebDreamer achieves substantial performance improvements over reactive baselines. It is competitive, while being 4-5 times more efficient, with tree search in sandbox environments (VisualWebArena) and also works effectively on real-world websites (Online-Mind2Web and Mind2Web-Live). Furthermore, our trained world model, Dreamer-7B, performs comparable to GPT-4o, highlighting the potential of specialized world models for efficient and effective planning in complex web environments.

comment: 22 pages, 11 figures, 6 tables

♻ ☆ ZETA: Leveraging Z-order Curves for Efficient Top-k Attention ICLR

Qiuhao Zeng, Jerry Huang, Peng Lu, Gezheng Xu, Boxing Chen, Charles Ling, Boyu Wang

Over recent years, the Transformer has become a fundamental building block for sequence modeling architectures. Yet at its core is the use of self-attention, whose memory and computational cost grow quadratically with the sequence length $N$, rendering it prohibitively expensive for long sequences. A promising approach is top-$k$ attention, which selects only the $k$ most relevant tokens and achieves performance comparable to vanilla self-attention while significantly reducing space and computational demands. However, causal masks require the current query token to only attend to past tokens, preventing the existing top-$k$ attention method from efficiently searching for the most relevant tokens in parallel, thereby limiting training efficiency. In this work, we propose ZETA, leveraging \textbf{Z}-Order Curves for \textbf{E}fficient \textbf{T}op-$k$ \textbf{A}ttention, to enable parallel querying of past tokens for entire sequences. % in both space and time complexity of $\mathcal{O}(N \log N)$. We first theoretically show that the choice of key and query dimensions involves a trade-off between the curse of dimensionality and the preservation of relative distances after projection. In light of this insight, we propose reducing the dimensionality of keys and queries in contrast to values and further leverage $Z$-order curves to map low-dimensional keys and queries into \emph{one}-dimensional space, which permits parallel sorting, thereby largely improving the efficiency for top-$k$ token selection. Experimental results demonstrate that ZETA matches the performance of standard attention on the synthetic \textsc{Multi-Query Associative Recall} task and outperforms attention and its variants on \textsc{Long Range Arena} and \textsc{WikiText-103} language modeling.

comment: 25 pages, 4 figures, accepted in International Conference on Learning Representations (ICLR) 2025

♻ ☆ Sparse identification of nonlinear dynamics and Koopman operators with Shallow Recurrent Decoder Networks

Mars Liyao Gao, Jan P. Williams, J. Nathan Kutz

Modeling real-world spatio-temporal data is exceptionally difficult due to inherent high dimensionality, measurement noise, partial observations, and often expensive data collection procedures. In this paper, we present Sparse Identification of Nonlinear Dynamics with SHallow REcurrent Decoder networks (SINDy-SHRED), a method to jointly solve the sensing and model identification problems with simple implementation, efficient computation, and robust performance. SINDy-SHRED uses Gated Recurrent Units to model the temporal sequence of sparse sensor measurements along with a shallow decoder network to reconstruct the full spatio-temporal field from the latent state space. Our algorithm introduces a SINDy-based regularization for which the latent space progressively converges to a SINDy-class functional, provided the projection remains within the set. In restricting SINDy to a linear model, a Koopman-SHRED model is generated. SINDy-SHRED (i) learns a symbolic and interpretable generative model of a parsimonious and low-dimensional latent space for the complex spatio-temporal dynamics, (ii) discovers new physics models even for well-known physical systems, (iii) achieves provably robust convergence with an observed globally convex loss landscape, and (iv) achieves superior accuracy, data efficiency, and training time, all with fewer model parameters. We conduct systematic experimental studies on PDE data such as turbulent flows, real-world sensor measurements for sea surface temperature, and direct video data. The interpretable SINDy and Koopman models of latent state dynamics enable stable and accurate long-term video predictions, outperforming all current baseline deep learning models in accuracy, training time, and data requirements, including Convolutional LSTM, PredRNN, ResNet, and SimVP.

♻ ☆ Diffusion State-Guided Projected Gradient for Inverse Problems ICLR 2025

Rayhan Zirvi, Bahareh Tolooshams, Anima Anandkumar

Recent advancements in diffusion models have been effective in learning data priors for solving inverse problems. They leverage diffusion sampling steps for inducing a data prior while using a measurement guidance gradient at each step to impose data consistency. For general inverse problems, approximations are needed when an unconditionally trained diffusion model is used since the measurement likelihood is intractable, leading to inaccurate posterior sampling. In other words, due to their approximations, these methods fail to preserve the generation process on the data manifold defined by the diffusion prior, leading to artifacts in applications such as image restoration. To enhance the performance and robustness of diffusion models in solving inverse problems, we propose Diffusion State-Guided Projected Gradient (DiffStateGrad), which projects the measurement gradient onto a subspace that is a low-rank approximation of an intermediate state of the diffusion process. DiffStateGrad, as a module, can be added to a wide range of diffusion-based inverse solvers to improve the preservation of the diffusion process on the prior manifold and filter out artifact-inducing components. We highlight that DiffStateGrad improves the robustness of diffusion models in terms of the choice of measurement guidance step size and noise while improving the worst-case performance. Finally, we demonstrate that DiffStateGrad improves upon the state-of-the-art on linear and nonlinear image restoration inverse problems. Our code is available at https://github.com/Anima-Lab/DiffStateGrad.

comment: Published as a conference paper at ICLR 2025. RZ and BT have equal contributions

♻ ☆ CodingTeachLLM: Empowering LLM's Coding Ability via AST Prior Knowledge

Zhangquan Chen, Chunjiang Liu, Haobin Duan

In this paper, we introduce CodingTeachLLM, a large language model (LLM) designed for coding teaching. Specially, we aim to enhance the coding ability of LLM and lead it to better teaching mode in education context. Thus, we propose an end-to-end prior-based three-phases supervised fine-tuned model, which is proved more competitive than traditional fine-tuning method. More specifically, our model realizes the structural disassembly and incremental guided output of educational knowledge. To this end, we robustify data classification of three types via a sampler and overlap estimation neural network, and inject the preprocessing datasets into pre-trained model in three batches for LORA fine-tuning. Then, we design a prior module couples system prompt, vector databases, and abstract syntax tree task segmentation. Finally, the compression method and regularization constraint are applied to the prior-based fine-tuned model, followed by text filter at the output end to obtain incremental guided results. Our model represents the first research effort to truly embody the tutor role with the features of abundant educational knowledge, step-by-step incremental guided outputs and non-disclosure of answers. Extensive experiments report that our model also achieves state-of-the-art in code abilities compared to open-source models, reaching an impressive 75.10% on the HumanEval (@pass 1) benchmark. Additionally, our model maintains strong conversational capabilities, with the 13B quantized version achieving scores of 56.34, 50.60, and 45.27 respectively on the MMLU, C-Eval, and AGIEval (5 shot) dialogue evaluation benchmarks.

comment: 9 pages, 2 figures

♻ ☆ VideoMind: A Chain-of-LoRA Agent for Long Video Reasoning

Ye Liu, Kevin Qinghong Lin, Chang Wen Chen, Mike Zheng Shou

Videos, with their unique temporal dimension, demand precise grounded understanding, where answers are directly linked to visual, interpretable evidence. Despite significant breakthroughs in reasoning capabilities within Large Language Models, multi-modal reasoning - especially for videos - remains unexplored. In this work, we introduce VideoMind, a novel video-language agent designed for temporal-grounded video understanding. VideoMind incorporates two key innovations: (i) We identify essential capabilities for video temporal reasoning and develop a role-based agentic workflow, including a planner for coordinating different roles, a grounder for temporal localization, a verifier to assess temporal interval accuracy, and an answerer for question-answering. (ii) To efficiently integrate these diverse roles, we propose a novel Chain-of-LoRA strategy, enabling seamless role-switching via lightweight LoRA adaptors while avoiding the overhead of multiple models, thus balancing efficiency and flexibility. Extensive experiments on 14 public benchmarks, including 3 on grounded video question-answering (Grounded VideoQA), 6 on video temporal grounding (VTG), and 5 on general video question-answering (VideoQA), verify that our agent achieves state-of-the-art performance on diverse video understanding tasks, underscoring its effectiveness in advancing video agent and long-form temporal reasoning.

comment: Project Page: https://videomind.github.io/

♻ ☆ Without Paired Labeled Data: An End-to-End Self-Supervised Paradigm for UAV-View Geo-Localization

Zhongwei Chen, Zhao-Xu Yang, Hai-Jun Rong

UAV-View Geo-Localization (UVGL) aims to achieve accurate localization of unmanned aerial vehicles (UAVs) by retrieving the most relevant GPS-tagged satellite images. However, existing methods heavily rely on pre-paired UAV-satellite images for supervised learning. Such dependency not only incurs high annotation costs but also severely limits scalability and practical deployment in open-world UVGL scenarios. To address these limitations, we propose an end-to-end self-supervised UVGL method. Our method leverages a shallow backbone network to extract initial features, employs clustering to generate pseudo labels, and adopts a dual-path contrastive learning architecture to learn discriminative intra-view representations. Furthermore, our method incorporates two core modules, the dynamic hierarchical memory learning module and the information consistency evolution learning module. The dynamic hierarchical memory learning module combines short-term and long-term memory to enhance intra-view feature consistency and discriminability. Meanwhile, the information consistency evolution learning module leverages a neighborhood-driven dynamic constraint mechanism to systematically capture implicit cross-view semantic correlations, thereby improving cross-view feature alignment. To further stabilize and strengthen the self-supervised training process, a pseudo-label enhancement strategy is introduced, which refines the quality of pseudo supervision. Our method ultimately constructs a unified cross-view feature representation space under self-supervised settings. Extensive experiments on three public benchmark datasets demonstrate that the proposed method consistently outperforms existing self-supervised methods and even surpasses several state-of-the-art supervised methods. Our code is available at https://github.com/ISChenawei/DMNIL.

♻ ☆ Data-Free Group-Wise Fully Quantized Winograd Convolution via Learnable Scales CVPR 2025

Shuokai Pan, Gerti Tuzi, Sudarshan Sreeram, Dibakar Gope

Despite the revolutionary breakthroughs of large-scale text-to-image diffusion models for complex vision and downstream tasks, their extremely high computational and storage costs limit their usability. Quantization of diffusion models has been explored in recent works to reduce compute costs and memory bandwidth usage. To further improve inference time, fast convolution algorithms such as Winograd can be used for convolution layers, which account for a significant portion of computations in diffusion models. However, the significant quality loss of fully quantized Winograd using existing coarser-grained post-training quantization methods, combined with the complexity and cost of finetuning the Winograd transformation matrices for such large models to recover quality, makes them unsuitable for large-scale foundation models. Motivated by the presence of a large range of values in them, we investigate the impact of finer-grained group-wise quantization in quantizing diffusion models. While group-wise quantization can largely handle the fully quantized Winograd convolution, it struggles to deal with the large distribution imbalance in a sizable portion of the Winograd domain computation. To reduce range differences in the Winograd domain, we propose finetuning only the scale parameters of the Winograd transform matrices without using any domain-specific training data. Because our method does not depend on any training data, the generalization performance of quantized diffusion models is safely guaranteed. For text-to-image generation task, the 8-bit fully-quantized diffusion model with Winograd provides near-lossless quality (FID and CLIP scores) in comparison to the full-precision model. For image classification, our method outperforms the state-of-the-art Winograd PTQ method by 1.62% and 2.56% in top-1 ImageNet accuracy on ResNet18 and ResNet-34, respectively, with Winograd F(6, 3).

comment: Accepted by CVPR 2025

♻ ☆ DC-SGD: Differentially Private SGD with Dynamic Clipping through Gradient Norm Distribution Estimation

Chengkun Wei, Weixian Li, Chen Gong, Wenzhi Chen

Differentially Private Stochastic Gradient Descent (DP-SGD) is a widely adopted technique for privacy-preserving deep learning. A critical challenge in DP-SGD is selecting the optimal clipping threshold C, which involves balancing the trade-off between clipping bias and noise magnitude, incurring substantial privacy and computing overhead during hyperparameter tuning. In this paper, we propose Dynamic Clipping DP-SGD (DC-SGD), a framework that leverages differentially private histograms to estimate gradient norm distributions and dynamically adjust the clipping threshold C. Our framework includes two novel mechanisms: DC-SGD-P and DC-SGD-E. DC-SGD-P adjusts the clipping threshold based on a percentile of gradient norms, while DC-SGD-E minimizes the expected squared error of gradients to optimize C. These dynamic adjustments significantly reduce the burden of hyperparameter tuning C. The extensive experiments on various deep learning tasks, including image classification and natural language processing, show that our proposed dynamic algorithms achieve up to 9 times acceleration on hyperparameter tuning than DP-SGD. And DC-SGD-E can achieve an accuracy improvement of 10.62% on CIFAR10 than DP-SGD under the same privacy budget of hyperparameter tuning. We conduct rigorous theoretical privacy and convergence analyses, showing that our methods seamlessly integrate with the Adam optimizer. Our results highlight the robust performance and efficiency of DC-SGD, offering a practical solution for differentially private deep learning with reduced computational overhead and enhanced privacy guarantees.

comment: Accepted at IEEE Transactions on Information Forensics & Security

♻ ☆ BounTCHA: A CAPTCHA Utilizing Boundary Identification in Guided Generative AI-extended Videos

Lehao Lin, Ke Wang, Maha Abdallah, Wei Cai

In recent years, the rapid development of artificial intelligence (AI) especially multi-modal Large Language Models (MLLMs), has enabled it to understand text, images, videos, and other multimedia data, allowing AI systems to execute various tasks based on human-provided prompts. However, AI-powered bots have increasingly been able to bypass most existing CAPTCHA systems, posing significant security threats to web applications. This makes the design of new CAPTCHA mechanisms an urgent priority. We observe that humans are highly sensitive to shifts and abrupt changes in videos, while current AI systems still struggle to comprehend and respond to such situations effectively. Based on this observation, we design and implement BounTCHA, a CAPTCHA mechanism that leverages human perception of boundaries in video transitions and disruptions. By utilizing generative AI's capability to extend original videos with prompts, we introduce unexpected twists and changes to create a pipeline for generating guided short videos for CAPTCHA purposes. We develop a prototype and conduct experiments to collect data on humans' time biases in boundary identification. This data serves as a basis for distinguishing between human users and bots. Additionally, we perform a detailed security analysis of BounTCHA, demonstrating its resilience against various types of attacks. We hope that BounTCHA will act as a robust defense, safeguarding millions of web applications in the AI-driven era.

comment: 22 pages, 15 figures; references added, typos corrected, new keyword "guided" added, new experimental data and related results updated; new keyword "Generative AI" added for clarity

♻ ☆ Visual Acoustic Fields

Yuelei Li, Hyunjin Kim, Fangneng Zhan, Ri-Zhao Qiu, Mazeyu Ji, Xiaojun Shan, Xueyan Zou, Paul Liang, Hanspeter Pfister, Xiaolong Wang

Objects produce different sounds when hit, and humans can intuitively infer how an object might sound based on its appearance and material properties. Inspired by this intuition, we propose Visual Acoustic Fields, a framework that bridges hitting sounds and visual signals within a 3D space using 3D Gaussian Splatting (3DGS). Our approach features two key modules: sound generation and sound localization. The sound generation module leverages a conditional diffusion model, which takes multiscale features rendered from a feature-augmented 3DGS to generate realistic hitting sounds. Meanwhile, the sound localization module enables querying the 3D scene, represented by the feature-augmented 3DGS, to localize hitting positions based on the sound sources. To support this framework, we introduce a novel pipeline for collecting scene-level visual-sound sample pairs, achieving alignment between captured images, impact locations, and corresponding sounds. To the best of our knowledge, this is the first dataset to connect visual and acoustic signals in a 3D context. Extensive experiments on our dataset demonstrate the effectiveness of Visual Acoustic Fields in generating plausible impact sounds and accurately localizing impact sources. Our project page is at https://yuelei0428.github.io/projects/Visual-Acoustic-Fields/.

♻ ☆ Astrea: A MOE-based Visual Understanding Model with Progressive Alignment

Xiaoda Yang, JunYu Lu, Hongshun Qiu, Sijing Li, Hao Li, Shengpeng Ji, Xudong Tang, Jiayang Xu, Jiaqi Duan, Ziyue Jiang, Cong Lin, Sihang Cai, Zejian Xie, Zhuoyang Song, Songxin Zhang

Vision-Language Models (VLMs) based on Mixture-of-Experts (MoE) architectures have emerged as a pivotal paradigm in multimodal understanding, offering a powerful framework for integrating visual and linguistic information. However, the increasing complexity and diversity of tasks present significant challenges in coordinating load balancing across heterogeneous visual experts, where optimizing one specialist's performance often compromises others' capabilities. To address task heterogeneity and expert load imbalance, we propose Astrea, a novel multi-expert collaborative VLM architecture based on progressive pre-alignment. Astrea introduces three key innovations: 1) A heterogeneous expert coordination mechanism that integrates four specialized models (detection, segmentation, classification, captioning) into a comprehensive expert matrix covering essential visual comprehension elements; 2) A dynamic knowledge fusion strategy featuring progressive pre-alignment to harmonize experts within the VLM latent space through contrastive learning, complemented by probabilistically activated stochastic residual connections to preserve knowledge continuity; 3) An enhanced optimization framework utilizing momentum contrastive learning for long-range dependency modeling and adaptive weight allocators for real-time expert contribution calibration. Extensive evaluations across 12 benchmark tasks spanning VQA, image captioning, and cross-modal retrieval demonstrate Astrea's superiority over state-of-the-art models, achieving an average performance gain of +4.7\%. This study provides the first empirical demonstration that progressive pre-alignment strategies enable VLMs to overcome task heterogeneity limitations, establishing new methodological foundations for developing general-purpose multimodal agents.

♻ ☆ Learned Image Compression and Restoration for Digital Pathology

SeonYeong Lee, EonSeung Seong, DongEon Lee, SiYeoul Lee, Yubin Cho, Chunsu Park, Seonho Kim, MinKyung Seo, YoungSin Ko, MinWoo Kim

Digital pathology images play a crucial role in medical diagnostics, but their ultra-high resolution and large file sizes pose significant challenges for storage, transmission, and real-time visualization. To address these issues, we propose CLERIC, a novel deep learning-based image compression framework designed specifically for whole slide images (WSIs). CLERIC integrates a learnable lifting scheme and advanced convolutional techniques to enhance compression efficiency while preserving critical pathological details. Our framework employs a lifting-scheme transform in the analysis stage to decompose images into low- and high-frequency components, enabling more structured latent representations. These components are processed through parallel encoders incorporating Deformable Residual Blocks (DRB) and Recurrent Residual Blocks (R2B) to improve feature extraction and spatial adaptability. The synthesis stage applies an inverse lifting transform for effective image reconstruction, ensuring high-fidelity restoration of fine-grained tissue structures. We evaluate CLERIC on a digital pathology image dataset and compare its performance against state-of-the-art learned image compression (LIC) models. Experimental results demonstrate that CLERIC achieves superior rate-distortion (RD) performance, significantly reducing storage requirements while maintaining high diagnostic image quality. Our study highlights the potential of deep learning-based compression in digital pathology, facilitating efficient data management and long-term storage while ensuring seamless integration into clinical workflows and AI-assisted diagnostic systems. Code and models are available at: https://github.com/pnu-amilab/CLERIC.

♻ ☆ Non-Determinism of "Deterministic" LLM Settings

Berk Atil, Sarp Aykent, Alexa Chittams, Lisheng Fu, Rebecca J. Passonneau, Evan Radcliffe, Guru Rajan Rajagopal, Adam Sloan, Tomasz Tudrej, Ferhan Ture, Zhe Wu, Lixinyu Xu, Breck Baldwin

LLM (large language model) practitioners commonly notice that outputs can vary for the same inputs under settings expected to be deterministic. Yet the questions of how pervasive this is, and with what impact on results, have not to our knowledge been systematically investigated. We investigate non-determinism in five LLMs configured to be deterministic when applied to eight common tasks in across 10 runs, in both zero-shot and few-shot settings. We see accuracy variations up to 15% across naturally occurring runs with a gap of best possible performance to worst possible performance up to 70%. In fact, none of the LLMs consistently delivers repeatable accuracy across all tasks, much less identical output strings. Sharing preliminary results with insiders has revealed that non-determinism perhaps essential to the efficient use of compute resources via co-mingled data in input buffers so this issue is not going away anytime soon. To better quantify our observations, we introduce metrics focused on quantifying determinism, TARr@N for the total agreement rate at N runs over raw output, and TARa@N for total agreement rate of parsed-out answers. Our code and data are publicly available at http://github.com/REDACTED.

♻ ☆ WaveFormer: A 3D Transformer with Wavelet-Driven Feature Representation for Efficient Medical Image Segmentation

Md Mahfuz Al Hasan, Mahdi Zaman, Abdul Jawad, Alberto Santamaria-Pang, Ho Hin Lee, Ivan Tarapov, Kyle See, Md Shah Imran, Antika Roy, Yaser Pourmohammadi Fallah, Navid Asadizanjani, Reza Forghani

Transformer-based architectures have advanced medical image analysis by effectively modeling long-range dependencies, yet they often struggle in 3D settings due to substantial memory overhead and insufficient capture of fine-grained local features. We address these limitations with WaveFormer, a novel 3D-transformer that: i) leverages the fundamental frequency-domain properties of features for contextual representation, and ii) is inspired by the top-down mechanism of the human visual recognition system, making it a biologically motivated architecture. By employing discrete wavelet transformations (DWT) at multiple scales, WaveFormer preserves both global context and high-frequency details while replacing heavy upsampling layers with efficient wavelet-based summarization and reconstruction. This significantly reduces the number of parameters, which is critical for real-world deployment where computational resources and training times are constrained. Furthermore, the model is generic and easily adaptable to diverse applications. Evaluations on BraTS2023, FLARE2021, and KiTS2023 demonstrate performance on par with state-of-the-art methods while offering substantially lower computational complexity.

♻ ☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems

Ooha Lakkadi Reddy

This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.

comment: 106 pages (42 main text, 6 references, 58 appendices). 21 figures, 4 tables in main text; 106 figures, 8 tables total. Code: https://github.com/oohalakkadi/ivc2tyc. Undergraduate thesis at Duke Kunshan University. Accepted for presentation at the 52nd International Conference for Computer Applications & Quantitative Methods in Archaeology (CAA 2025), Athens, Greece

♻ ☆ Towards shutdownable agents via stochastic choice

Elliott Thornley, Alexander Roman, Christos Ziakas, Leyton Ho, Louis Thomson

The Incomplete Preferences Proposal (IPP) is an idea for ensuring that advanced artificial agents never resist shutdown. A key part of the IPP is using a novel `Discounted Reward for Same-Length Trajectories (DReST)' reward function to train agents to (1) pursue goals effectively conditional on each trajectory-length (be `USEFUL'), and (2) choose stochastically between different trajectory-lengths (be `NEUTRAL' about trajectory-lengths). In this paper, we propose evaluation metrics for USEFULNESS and NEUTRALITY. We use a DReST reward function to train simple agents to navigate gridworlds, and we find that these agents learn to be USEFUL and NEUTRAL. Our results thus provide some initial evidence that DReST reward functions could train advanced agents to be USEFUL and NEUTRAL. Our theoretical work suggests that these agents would be useful and shutdownable.

♻ ☆ Provably-Safe Neural Network Training Using Hybrid Zonotope Reachability Analysis

Long Kiu Chung, Shreyas Kousik

Even though neural networks are being increasingly deployed in safety-critical control applications, it remains difficult to enforce constraints on their output, meaning that it is hard to guarantee safety in such settings. While many existing methods seek to verify a neural network's satisfaction of safety constraints, few address how to correct an unsafe network. The handful of works that extract a training signal from verification cannot handle non-convex sets, and are either conservative or slow. To begin addressing these challenges, this work proposes a neural network training method that can encourage the exact image of a non-convex input set for a neural network with rectified linear unit (ReLU) nonlinearities to avoid a non-convex unsafe region. This is accomplished by reachability analysis with scaled hybrid zonotopes, a modification of the existing hybrid zonotope set representation that enables parameterized scaling of non-convex polytopic sets with a differentiable collision check via mixed-integer linear programs (MILPs). The proposed method was shown to be effective and fast for networks with up to 240 neurons, with the computational complexity dominated by inverse operations on matrices that scale linearly in size with the number of neurons and complexity of input and unsafe sets. We demonstrate the practicality of our method by training a forward-invariant neural network controller for a non-convex input set to an affine system, as well as generating safe reach-avoid plans for a black-box dynamical system.

comment: 8 pages, 3 figures

♻ ☆ View-Invariant Pixelwise Anomaly Detection in Multi-object Scenes with Adaptive View Synthesis

Subin Varghese, Vedhus Hoskere

Visual anomaly detection in the built environment is a valuable tool for applications such as infrastructure assessment, construction monitoring, security surveillance, and urban planning. Anomaly detection approaches are typically unsupervised and work by detecting deviations from an expected state where no assumptions are made exact type of deviation. Unsupervised pixel-level anomaly detection methods have been developed to successfully recognize and segment anomalies; however, existing techniques are designed for industrial settings with a fixed camera position. In the built environment, images are periodically captured by a camera operated manually or mounted on aerial or ground vehicles. The camera pose between successive collections may vary widely voiding a fundamental assumption in existing anomaly detection approaches. To address this gap, we introduce the problem of Scene Anomaly Detection (Scene AD), where the goal is to detect anomalies from two sets of images: one set without anomalies and one set that may or may not contain anomalies. No labeled semantic segmentation data are provided for training. We propose a novel network, OmniAD, to tackle Scene AD by refining the reverse distillation anomaly detection method, leading to a 40\% improvement in pixel-level anomaly detection. Additionally, we introduce two new data augmentation strategies that leverage novel view synthesis and camera localization to enhance generalization. We evaluate our approach both qualitatively and quantitatively on a new dataset, ToyCity the first Scene AD dataset featuring multiple objects as well as on the established single object centric dataset, MAD. Our method demonstrates marked improvement over baseline approaches, paving the way for robust anomaly detection in scenes with real-world camera pose variations commonly observed in the built environment. https://drags99.github.io/OmniAD/

♻ ☆ Heterogeneous bimodal attention fusion for speech emotion recognition

Jiachen Luo, Huy Phan, Lin Wang, Joshua Reiss

Multi-modal emotion recognition in conversations is a challenging problem due to the complex and complementary interactions between different modalities. Audio and textual cues are particularly important for understanding emotions from a human perspective. Most existing studies focus on exploring interactions between audio and text modalities at the same representation level. However, a critical issue is often overlooked: the heterogeneous modality gap between low-level audio representations and high-level text representations. To address this problem, we propose a novel framework called Heterogeneous Bimodal Attention Fusion (HBAF) for multi-level multi-modal interaction in conversational emotion recognition. The proposed method comprises three key modules: the uni-modal representation module, the multi-modal fusion module, and the inter-modal contrastive learning module. The uni-modal representation module incorporates contextual content into low-level audio representations to bridge the heterogeneous multi-modal gap, enabling more effective fusion. The multi-modal fusion module uses dynamic bimodal attention and a dynamic gating mechanism to filter incorrect cross-modal relationships and fully exploit both intra-modal and inter-modal interactions. Finally, the inter-modal contrastive learning module captures complex absolute and relative interactions between audio and text modalities. Experiments on the MELD and IEMOCAP datasets demonstrate that the proposed HBAF method outperforms existing state-of-the-art baselines.

♻ ☆ How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approach

Ayeong Lee, Ethan Che, Tianyi Peng

Chain-of-thought prompting has emerged as a powerful technique for enabling large language models (LLMs) to solve complex reasoning tasks. However, these reasoning chains can be verbose, raising concerns about efficiency. In response, recent works have sought to decrease response lengths through simple prompting strategies (e.g. 'be concise'). In this work, we conduct the first systematic study of the relationship between reasoning length and model performance across a diverse range of compression instructions (e.g. 'use 10 words or less' or 'remove all punctuation'). In doing so, we discover a universal tradeoff between reasoning length and accuracy that persists across even very distinct reasoning chains. We demonstrate that this tradeoff emerges from a sharp threshold behavior at the question level: each task has an intrinsic 'token complexity' - a minimal number of tokens required for successful problem-solving. We show how token complexity enables us to compute information-theoretic limits on the accuracy-compression tradeoff, and find that prompt-based compression strategies operate far from these theoretical limits. This suggests there may be significant room for improvement and our framework provides a benchmark to help researchers evaluate progress in reasoning efficiency. Our work also highlights the importance of adaptive compression -- giving shorter responses for easier questions -- and we show that token complexity is a useful tool for measuring this capability.

♻ ☆ Independent and Decentralized Learning in Markov Potential Games

Chinmay Maheshwari, Manxi Wu, Druv Pai, Shankar Sastry

We study a multi-agent reinforcement learning dynamics, and analyze its asymptotic behavior in infinite-horizon discounted Markov potential games. We focus on the independent and decentralized setting, where players do not know the game parameters, and cannot communicate or coordinate. In each stage, players update their estimate of Q-function that evaluates their total contingent payoff based on the realized one-stage reward in an asynchronous manner. Then, players independently update their policies by incorporating an optimal one-stage deviation strategy based on the estimated Q-function. Inspired by the actor-critic algorithm in single-agent reinforcement learning, a key feature of our learning dynamics is that agents update their Q-function estimates at a faster timescale than the policies. Leveraging tools from two-timescale asynchronous stochastic approximation theory, we characterize the convergent set of learning dynamics.

comment: 43 pages, 1 figure

♻ ☆ Convergence of Decentralized Actor-Critic Algorithm in General-sum Markov Games

Chinmay Maheshwari, Manxi Wu, Shankar Sastry

Markov games provide a powerful framework for modeling strategic multi-agent interactions in dynamic environments. Traditionally, convergence properties of decentralized learning algorithms in these settings have been established only for special cases, such as Markov zero-sum and potential games, which do not fully capture real-world interactions. In this paper, we address this gap by studying the asymptotic properties of learning algorithms in general-sum Markov games. In particular, we focus on a decentralized algorithm where each agent adopts an actor-critic learning dynamic with asynchronous step sizes. This decentralized approach enables agents to operate independently, without requiring knowledge of others' strategies or payoffs. We introduce the concept of a Markov Near-Potential Function (MNPF) and demonstrate that it serves as an approximate Lyapunov function for the policy updates in the decentralized learning dynamics, which allows us to characterize the convergent set of strategies. We further strengthen our result under specific regularity conditions and with finite Nash equilibria.

comment: 18 pages, 3 figure

♻ ☆ Process or Result? Manipulated Ending Tokens Can Mislead Reasoning LLMs to Ignore the Correct Reasoning Steps

Yu Cui, Bryan Hooi, Yujun Cai, Yiwei Wang

Recent reasoning large language models (LLMs) have demonstrated remarkable improvements in mathematical reasoning capabilities through long Chain-of-Thought. The reasoning tokens of these models enable self-correction within reasoning chains, enhancing robustness. This motivates our exploration: how vulnerable are reasoning LLMs to subtle errors in their input reasoning chains? We introduce "Compromising Thought" (CPT), a vulnerability where models presented with reasoning tokens containing manipulated calculation results tend to ignore correct reasoning steps and adopt incorrect results instead. Through systematic evaluation across multiple reasoning LLMs, we design three increasingly explicit prompting methods to measure CPT resistance, revealing that models struggle significantly to identify and correct these manipulations. Notably, contrary to existing research suggesting structural alterations affect model performance more than content modifications, we find that local ending token manipulations have greater impact on reasoning outcomes than structural changes. Moreover, we discover a security vulnerability in DeepSeek-R1 where tampered reasoning tokens can trigger complete reasoning cessation. Our work enhances understanding of reasoning robustness and highlights security considerations for reasoning-intensive applications.

☆ UniOcc: A Unified Benchmark for Occupancy Forecasting and Prediction in Autonomous Driving

Yuping Wang, Xiangyu Huang, Xiaokang Sun, Mingxuan Yan, Shuo Xing, Zhengzhong Tu, Jiachen Li

We introduce UniOcc, a comprehensive, unified benchmark for occupancy forecasting (i.e., predicting future occupancies based on historical information) and current-frame occupancy prediction from camera images. UniOcc unifies data from multiple real-world datasets (i.e., nuScenes, Waymo) and high-fidelity driving simulators (i.e., CARLA, OpenCOOD), which provides 2D/3D occupancy labels with per-voxel flow annotations and support for cooperative autonomous driving. In terms of evaluation, unlike existing studies that rely on suboptimal pseudo labels for evaluation, UniOcc incorporates novel metrics that do not depend on ground-truth occupancy, enabling robust assessment of additional aspects of occupancy quality. Through extensive experiments on state-of-the-art models, we demonstrate that large-scale, diverse training data and explicit flow information significantly enhance occupancy prediction and forecasting performance.

comment: 14 pages; Dataset: https://huggingface.co/datasets/tasl-lab/uniocc; Code: https://github.com/tasl-lab/UniOcc

☆ Sim-and-Real Co-Training: A Simple Recipe for Vision-Based Robotic Manipulation

Abhiram Maddukuri, Zhenyu Jiang, Lawrence Yunliang Chen, Soroush Nasiriany, Yuqi Xie, Yu Fang, Wenqi Huang, Zu Wang, Zhenjia Xu, Nikita Chernyadev, Scott Reed, Ken Goldberg, Ajay Mandlekar, Linxi Fan, Yuke Zhu

Large real-world robot datasets hold great potential to train generalist robot models, but scaling real-world human data collection is time-consuming and resource-intensive. Simulation has great potential in supplementing large-scale data, especially with recent advances in generative AI and automated data generation tools that enable scalable creation of robot behavior datasets. However, training a policy solely in simulation and transferring it to the real world often demands substantial human effort to bridge the reality gap. A compelling alternative is to co-train the policy on a mixture of simulation and real-world datasets. Preliminary studies have recently shown this strategy to substantially improve the performance of a policy over one trained on a limited amount of real-world data. Nonetheless, the community lacks a systematic understanding of sim-and-real co-training and what it takes to reap the benefits of simulation data for real-robot learning. This work presents a simple yet effective recipe for utilizing simulation data to solve vision-based robotic manipulation tasks. We derive this recipe from comprehensive experiments that validate the co-training strategy on various simulation and real-world datasets. Using two domains--a robot arm and a humanoid--across diverse tasks, we demonstrate that simulation data can enhance real-world task performance by an average of 38%, even with notable differences between the simulation and real-world data. Videos and additional results can be found at https://co-training.github.io/

comment: Project website: https://co-training.github.io/

☆ Pro-Routing: Proactive Routing of Autonomous Multi-Capacity Robots for Pickup-and-Delivery Tasks

Daniel Garces, Stephanie Gil

We consider a multi-robot setting, where we have a fleet of multi-capacity autonomous robots that must service spatially distributed pickup-and-delivery requests with fixed maximum wait times. Requests can be either scheduled ahead of time or they can enter the system in real-time. In this setting, stability for a routing policy is defined as the cost of the policy being uniformly bounded over time. Most previous work either solve the problem offline to theoretically maintain stability or they consider dynamically arriving requests at the expense of the theoretical guarantees on stability. In this paper, we aim to bridge this gap by proposing a novel proactive rollout-based routing framework that adapts to real-time demand while still provably maintaining the stability of the learned routing policy. We derive provable stability guarantees for our method by proposing a fleet sizing algorithm that obtains a sufficiently large fleet that ensures stability by construction. To validate our theoretical results, we consider a case study on real ride requests for Harvard's evening Van System. We also evaluate the performance of our framework using the currently deployed smaller fleet size. In this smaller setup, we compare against the currently deployed routing algorithm, greedy heuristics, and Monte-Carlo-Tree-Search-based algorithms. Our empirical results show that our framework maintains stability when we use the sufficiently large fleet size found in our theoretical results. For the smaller currently deployed fleet size, our method services 6% more requests than the closest baseline while reducing median passenger wait times by 33%.

comment: 25 pages, 7 figures, and 1 table

☆ AutoEval: Autonomous Evaluation of Generalist Robot Manipulation Policies in the Real World

Zhiyuan Zhou, Pranav Atreya, You Liang Tan, Karl Pertsch, Sergey Levine

Scalable and reproducible policy evaluation has been a long-standing challenge in robot learning. Evaluations are critical to assess progress and build better policies, but evaluation in the real world, especially at a scale that would provide statistically reliable results, is costly in terms of human time and hard to obtain. Evaluation of increasingly generalist robot policies requires an increasingly diverse repertoire of evaluation environments, making the evaluation bottleneck even more pronounced. To make real-world evaluation of robotic policies more practical, we propose AutoEval, a system to autonomously evaluate generalist robot policies around the clock with minimal human intervention. Users interact with AutoEval by submitting evaluation jobs to the AutoEval queue, much like how software jobs are submitted with a cluster scheduling system, and AutoEval will schedule the policies for evaluation within a framework supplying automatic success detection and automatic scene resets. We show that AutoEval can nearly fully eliminate human involvement in the evaluation process, permitting around the clock evaluations, and the evaluation results correspond closely to ground truth evaluations conducted by hand. To facilitate the evaluation of generalist policies in the robotics community, we provide public access to multiple AutoEval scenes in the popular BridgeData robot setup with WidowX robot arms. In the future, we hope that AutoEval scenes can be set up across institutions to form a diverse and distributed evaluation network.

☆ Pseudo-Random UAV Test Generation Using Low-Fidelity Path Simulator

Anas Shrinah, Kerstin Eder

Simulation-based testing provides a safe and cost-effective environment for verifying the safety of Uncrewed Aerial Vehicles (UAVs). However, simulation can be resource-consuming, especially when High-Fidelity Simulators (HFS) are used. To optimise simulation resources, we propose a pseudo-random test generator that uses a Low-Fidelity Simulator (LFS) to estimate UAV flight paths. This work simplifies the PX4 autopilot HFS to develop a LFS, which operates one order of magnitude faster than the HFS.Test cases predicted to cause safety violations in the LFS are subsequently validated using the HFS.

☆ Reinforcement Learning for Safe Autonomous Two Device Navigation of Cerebral Vessels in Mechanical Thrombectomy

Harry Robertshaw, Benjamin Jackson, Jiaheng Wang, Hadi Sadati, Lennart Karstensen, Alejandro Granados, Thomas C Booth

Purpose: Autonomous systems in mechanical thrombectomy (MT) hold promise for reducing procedure times, minimizing radiation exposure, and enhancing patient safety. However, current reinforcement learning (RL) methods only reach the carotid arteries, are not generalizable to other patient vasculatures, and do not consider safety. We propose a safe dual-device RL algorithm that can navigate beyond the carotid arteries to cerebral vessels. Methods: We used the Simulation Open Framework Architecture to represent the intricacies of cerebral vessels, and a modified Soft Actor-Critic RL algorithm to learn, for the first time, the navigation of micro-catheters and micro-guidewires. We incorporate patient safety metrics into our reward function by integrating guidewire tip forces. Inverse RL is used with demonstrator data on 12 patient-specific vascular cases. Results: Our simulation demonstrates successful autonomous navigation within unseen cerebral vessels, achieving a 96% success rate, 7.0s procedure time, and 0.24 N mean forces, well below the proposed 1.5 N vessel rupture threshold. Conclusion: To the best of our knowledge, our proposed autonomous system for MT two-device navigation reaches cerebral vessels, considers safety, and is generalizable to unseen patient-specific cases for the first time. We envisage future work will extend the validation to vasculatures of different complexity and on in vitro models. While our contributions pave the way towards deploying agents in clinical settings, safety and trustworthiness will be crucial elements to consider when proposing new methodology.

☆ Graph Neural Network-Based Predictive Modeling for Robotic Plaster Printing

Diego Machain Rivera, Selen Ercan Jenny, Ping Hsun Tsai, Ena Lloret-Fritschi, Luis Salamanca, Fernando Perez-Cruz, Konstantinos E. Tatsis

This work proposes a Graph Neural Network (GNN) modeling approach to predict the resulting surface from a particle based fabrication process. The latter consists of spray-based printing of cementitious plaster on a wall and is facilitated with the use of a robotic arm. The predictions are computed using the robotic arm trajectory features, such as position, velocity and direction, as well as the printing process parameters. The proposed approach, based on a particle representation of the wall domain and the end effector, allows for the adoption of a graph-based solution. The GNN model consists of an encoder-processor-decoder architecture and is trained using data from laboratory tests, while the hyperparameters are optimized by means of a Bayesian scheme. The aim of this model is to act as a simulator of the printing process, and ultimately used for the generation of the robotic arm trajectory and the optimization of the printing parameters, towards the materialization of an autonomous plastering process. The performance of the proposed model is assessed in terms of the prediction error against unseen ground truth data, which shows its generality in varied scenarios, as well as in comparison with the performance of an existing benchmark model. The results demonstrate a significant improvement over the benchmark model, with notably better performance and enhanced error scaling across prediction steps.

☆ HACTS: a Human-As-Copilot Teleoperation System for Robot Learning

Zhiyuan Xu, Yinuo Zhao, Kun Wu, Ning Liu, Junjie Ji, Zhengping Che, Chi Harold Liu, Jian Tang

Teleoperation is essential for autonomous robot learning, especially in manipulation tasks that require human demonstrations or corrections. However, most existing systems only offer unilateral robot control and lack the ability to synchronize the robot's status with the teleoperation hardware, preventing real-time, flexible intervention. In this work, we introduce HACTS (Human-As-Copilot Teleoperation System), a novel system that establishes bilateral, real-time joint synchronization between a robot arm and teleoperation hardware. This simple yet effective feedback mechanism, akin to a steering wheel in autonomous vehicles, enables the human copilot to intervene seamlessly while collecting action-correction data for future learning. Implemented using 3D-printed components and low-cost, off-the-shelf motors, HACTS is both accessible and scalable. Our experiments show that HACTS significantly enhances performance in imitation learning (IL) and reinforcement learning (RL) tasks, boosting IL recovery capabilities and data efficiency, and facilitating human-in-the-loop RL. HACTS paves the way for more effective and interactive human-robot collaboration and data-collection, advancing the capabilities of robot manipulation.

☆ COSMO: Combination of Selective Memorization for Low-cost Vision-and-Language Navigation

Siqi Zhang, Yanyuan Qiao, Qunbo Wang, Zike Yan, Qi Wu, Zhihua Wei, Jing Liu

Vision-and-Language Navigation (VLN) tasks have gained prominence within artificial intelligence research due to their potential application in fields like home assistants. Many contemporary VLN approaches, while based on transformer architectures, have increasingly incorporated additional components such as external knowledge bases or map information to enhance performance. These additions, while boosting performance, also lead to larger models and increased computational costs. In this paper, to achieve both high performance and low computational costs, we propose a novel architecture with the COmbination of Selective MemOrization (COSMO). Specifically, COSMO integrates state-space modules and transformer modules, and incorporates two VLN-customized selective state space modules: the Round Selective Scan (RSS) and the Cross-modal Selective State Space Module (CS3). RSS facilitates comprehensive inter-modal interactions within a single scan, while the CS3 module adapts the selective state space module into a dual-stream architecture, thereby enhancing the acquisition of cross-modal interactions. Experimental validations on three mainstream VLN benchmarks, REVERIE, R2R, and R2R-CE, not only demonstrate competitive navigation performance of our model but also show a significant reduction in computational costs.

☆ Toward Anxiety-Reducing Pocket Robots for Children

Morten Roed Frederiksen, Kasper Støy, Maja Matarić

A common denominator for most therapy treatments for children who suffer from an anxiety disorder is daily practice routines to learn techniques needed to overcome anxiety. However, applying those techniques while experiencing anxiety can be highly challenging. This paper presents the design, implementation, and pilot study of a tactile hand-held pocket robot AffectaPocket, designed to work alongside therapy as a focus object to facilitate coping during an anxiety attack. The robot does not require daily practice to be used, has a small form factor, and has been designed for children 7 to 12 years old. The pocket robot works by sensing when it is being held and attempts to shift the child's focus by presenting them with a simple three-note rhythm-matching game. We conducted a pilot study of the pocket robot involving four children aged 7 to 10 years, and then a main study with 18 children aged 6 to 8 years; neither study involved children with anxiety. Both studies aimed to assess the reliability of the robot's sensor configuration, its design, and the effectiveness of the user tutorial. The results indicate that the morphology and sensor setup performed adequately and the tutorial process enabled the children to use the robot with little practice. This work demonstrates that the presented pocket robot could represent a step toward developing low-cost accessible technologies to help children suffering from anxiety disorders.

comment: 8 pages

☆ Learning 3D-Gaussian Simulators from RGB Videos

Mikel Zhobro, Andreas René Geist, Georg Martius

Learning physics simulations from video data requires maintaining spatial and temporal consistency, a challenge often addressed with strong inductive biases or ground-truth 3D information -- limiting scalability and generalization. We introduce 3DGSim, a 3D physics simulator that learns object dynamics end-to-end from multi-view RGB videos. It encodes images into a 3D Gaussian particle representation, propagates dynamics via a transformer, and renders frames using 3D Gaussian splatting. By jointly training inverse rendering with a dynamics transformer using a temporal encoding and merging layer, 3DGSimembeds physical properties into point-wise latent vectors without enforcing explicit connectivity constraints. This enables the model to capture diverse physical behaviors, from rigid to elastic and cloth-like interactions, along with realistic lighting effects that also generalize to unseen multi-body interactions and novel scene edits.

☆ SALT: A Flexible Semi-Automatic Labeling Tool for General LiDAR Point Clouds with Cross-Scene Adaptability and 4D Consistency

Yanbo Wang, Yongtao Chen, Chuan Cao, Tianchen Deng, Wentao Zhao, Jingchuan Wang, Weidong Chen

We propose a flexible Semi-Automatic Labeling Tool (SALT) for general LiDAR point clouds with cross-scene adaptability and 4D consistency. Unlike recent approaches that rely on camera distillation, SALT operates directly on raw LiDAR data, automatically generating pre-segmentation results. To achieve this, we propose a novel zero-shot learning paradigm, termed data alignment, which transforms LiDAR data into pseudo-images by aligning with the training distribution of vision foundation models. Additionally, we design a 4D-consistent prompting strategy and 4D non-maximum suppression module to enhance SAM2, ensuring high-quality, temporally consistent presegmentation. SALT surpasses the latest zero-shot methods by 18.4% PQ on SemanticKITTI and achieves nearly 40-50% of human annotator performance on our newly collected low-resolution LiDAR data and on combined data from three LiDAR types, significantly boosting annotation efficiency. We anticipate that SALT's open-sourcing will catalyze substantial expansion of current LiDAR datasets and lay the groundwork for the future development of LiDAR foundation models. Code is available at https://github.com/Cavendish518/SALT.

☆ A Reactive Framework for Whole-Body Motion Planning of Mobile Manipulators Combining Reinforcement Learning and SDF-Constrained Quadratic Programmi

Chenyu Zhang, Shiying Sun, Kuan Liu, Chuanbao Zhou, Xiaoguang Zhao, Min Tan, Yanlong Huang

As an important branch of embodied artificial intelligence, mobile manipulators are increasingly applied in intelligent services, but their redundant degrees of freedom also limit efficient motion planning in cluttered environments. To address this issue, this paper proposes a hybrid learning and optimization framework for reactive whole-body motion planning of mobile manipulators. We develop the Bayesian distributional soft actor-critic (Bayes-DSAC) algorithm to improve the quality of value estimation and the convergence performance of the learning. Additionally, we introduce a quadratic programming method constrained by the signed distance field to enhance the safety of the obstacle avoidance motion. We conduct experiments and make comparison with standard benchmark. The experimental results verify that our proposed framework significantly improves the efficiency of reactive whole-body motion planning, reduces the planning time, and improves the success rate of motion planning. Additionally, the proposed reinforcement learning method ensures a rapid learning process in the whole-body planning task. The novel framework allows mobile manipulators to adapt to complex environments more safely and efficiently.

☆ Video-based Traffic Light Recognition by Rockchip RV1126 for Autonomous Driving

Miao Fan, Xuxu Kong, Shengtong Xu, Haoyi Xiong, Xiangzeng Liu

Real-time traffic light recognition is fundamental for autonomous driving safety and navigation in urban environments. While existing approaches rely on single-frame analysis from onboard cameras, they struggle with complex scenarios involving occlusions and adverse lighting conditions. We present \textit{ViTLR}, a novel video-based end-to-end neural network that processes multiple consecutive frames to achieve robust traffic light detection and state classification. The architecture leverages a transformer-like design with convolutional self-attention modules, which is optimized specifically for deployment on the Rockchip RV1126 embedded platform. Extensive evaluations on two real-world datasets demonstrate that \textit{ViTLR} achieves state-of-the-art performance while maintaining real-time processing capabilities (>25 FPS) on RV1126's NPU. The system shows superior robustness across temporal stability, varying target distances, and challenging environmental conditions compared to existing single-frame approaches. We have successfully integrated \textit{ViTLR} into an ego-lane traffic light recognition system using HD maps for autonomous driving applications. The complete implementation, including source code and datasets, is made publicly available to facilitate further research in this domain.

comment: Accepted by IEEE IV'25

☆ A Benchmark for Vision-Centric HD Mapping by V2I Systems

Miao Fan, Shanshan Yu, Shengtong Xu, Kun Jiang, Haoyi Xiong, Xiangzeng Liu

Autonomous driving faces safety challenges due to a lack of global perspective and the semantic information of vectorized high-definition (HD) maps. Information from roadside cameras can greatly expand the map perception range through vehicle-to-infrastructure (V2I) communications. However, there is still no dataset from the real world available for the study on map vectorization onboard under the scenario of vehicle-infrastructure cooperation. To prosper the research on online HD mapping for Vehicle-Infrastructure Cooperative Autonomous Driving (VICAD), we release a real-world dataset, which contains collaborative camera frames from both vehicles and roadside infrastructures, and provides human annotations of HD map elements. We also present an end-to-end neural framework (i.e., V2I-HD) leveraging vision-centric V2I systems to construct vectorized maps. To reduce computation costs and further deploy V2I-HD on autonomous vehicles, we introduce a directionally decoupled self-attention mechanism to V2I-HD. Extensive experiments show that V2I-HD has superior performance in real-time inference speed, as tested by our real-world dataset. Abundant qualitative results also demonstrate stable and robust map construction quality with low cost in complex and various driving scenes. As a benchmark, both source codes and the dataset have been released at OneDrive for the purpose of further study.

comment: Accepted by IEEE IV'25

☆ MAER-Nav: Bidirectional Motion Learning Through Mirror-Augmented Experience Replay for Robot Navigation

Shanze Wang, Mingao Tan, Zhibo Yang, Biao Huang, Xiaoyu Shen, Hailong Huang, Wei Zhang

Deep Reinforcement Learning (DRL) based navigation methods have demonstrated promising results for mobile robots, but suffer from limited action flexibility in confined spaces. Conventional DRL approaches predominantly learn forward-motion policies, causing robots to become trapped in complex environments where backward maneuvers are necessary for recovery. This paper presents MAER-Nav (Mirror-Augmented Experience Replay for Robot Navigation), a novel framework that enables bidirectional motion learning without requiring explicit failure-driven hindsight experience replay or reward function modifications. Our approach integrates a mirror-augmented experience replay mechanism with curriculum learning to generate synthetic backward navigation experiences from successful trajectories. Experimental results in both simulation and real-world environments demonstrate that MAER-Nav significantly outperforms state-of-the-art methods while maintaining strong forward navigation capabilities. The framework effectively bridges the gap between the comprehensive action space utilization of traditional planning methods and the environmental adaptability of learning-based approaches, enabling robust navigation in scenarios where conventional DRL methods consistently fail.

comment: 8 pages, 8 figures

☆ Less is More: Contextual Sampling for Nonlinear Data-Enabled Predictive Control IROS 2025

Julius Beerwerth, Bassam Alrifaee

Data-enabled Predictive Control (DeePC) is a powerful data-driven approach for predictive control without requiring an explicit system model. However, its high computational cost limits its applicability to real-time robotic systems. For robotic applications such as motion planning and trajectory tracking, real-time control is crucial. Nonlinear DeePC either relies on large datasets or learning the nonlinearities to ensure predictive accuracy, leading to high computational complexity. This work introduces contextual sampling, a novel data selection strategy to handle nonlinearities for DeePC by dynamically selecting the most relevant data at each time step. By reducing the dataset size while preserving prediction accuracy, our method improves computational efficiency, of DeePC for real-time robotic applications. We validate our approach for autonomous vehicle motion planning. For a dataset size of 100 sub-trajectories, Contextual sampling DeePC reduces tracking error by 53.2 % compared to Leverage Score sampling. Additionally, Contextual sampling reduces max computation time by 87.2 % compared to using the full dataset of 491 sub-trajectories while achieving comparable tracking performance. These results highlight the potential of Contextual sampling to enable real-time, data-driven control for robotic systems.

comment: Submitted to IROS 2025 on March 1st

☆ ZeroMimic: Distilling Robotic Manipulation Skills from Web Videos ICRA 2025

Junyao Shi, Zhuolun Zhao, Tianyou Wang, Ian Pedroza, Amy Luo, Jie Wang, Jason Ma, Dinesh Jayaraman

Many recent advances in robotic manipulation have come through imitation learning, yet these rely largely on mimicking a particularly hard-to-acquire form of demonstrations: those collected on the same robot in the same room with the same objects as the trained policy must handle at test time. In contrast, large pre-recorded human video datasets demonstrating manipulation skills in-the-wild already exist, which contain valuable information for robots. Is it possible to distill a repository of useful robotic skill policies out of such data without any additional requirements on robot-specific demonstrations or exploration? We present the first such system ZeroMimic, that generates immediately deployable image goal-conditioned skill policies for several common categories of manipulation tasks (opening, closing, pouring, pick&place, cutting, and stirring) each capable of acting upon diverse objects and across diverse unseen task setups. ZeroMimic is carefully designed to exploit recent advances in semantic and geometric visual understanding of human videos, together with modern grasp affordance detectors and imitation policy classes. After training ZeroMimic on the popular EpicKitchens dataset of ego-centric human videos, we evaluate its out-of-the-box performance in varied real-world and simulated kitchen settings with two different robot embodiments, demonstrating its impressive abilities to handle these varied tasks. To enable plug-and-play reuse of ZeroMimic policies on other task setups and robots, we release software and policy checkpoints of our skill policies.

comment: ICRA 2025. Project website: https://zeromimic.github.io/

☆ GenSwarm: Scalable Multi-Robot Code-Policy Generation and Deployment via Language Models

Wenkang Ji, Huaben Chen, Mingyang Chen, Guobin Zhu, Lufeng Xu, Roderich Groß, Rui Zhou, Ming Cao, Shiyu Zhao

The development of control policies for multi-robot systems traditionally follows a complex and labor-intensive process, often lacking the flexibility to adapt to dynamic tasks. This has motivated research on methods to automatically create control policies. However, these methods require iterative processes of manually crafting and refining objective functions, thereby prolonging the development cycle. This work introduces \textit{GenSwarm}, an end-to-end system that leverages large language models to automatically generate and deploy control policies for multi-robot tasks based on simple user instructions in natural language. As a multi-language-agent system, GenSwarm achieves zero-shot learning, enabling rapid adaptation to altered or unseen tasks. The white-box nature of the code policies ensures strong reproducibility and interpretability. With its scalable software and hardware architectures, GenSwarm supports efficient policy deployment on both simulated and real-world multi-robot systems, realizing an instruction-to-execution end-to-end functionality that could prove valuable for robotics specialists and non-specialists alike.The code of the proposed GenSwarm system is available online: https://github.com/WindyLab/GenSwarm.

☆ Disambiguate Gripper State in Grasp-Based Tasks: Pseudo-Tactile as Feedback Enables Pure Simulation Learning IROS 2025

Yifei Yang, Lu Chen, Zherui Song, Yenan Chen, Wentao Sun, Zhongxiang Zhou, Rong Xiong, Yue Wang

Grasp-based manipulation tasks are fundamental to robots interacting with their environments, yet gripper state ambiguity significantly reduces the robustness of imitation learning policies for these tasks. Data-driven solutions face the challenge of high real-world data costs, while simulation data, despite its low costs, is limited by the sim-to-real gap. We identify the root cause of gripper state ambiguity as the lack of tactile feedback. To address this, we propose a novel approach employing pseudo-tactile as feedback, inspired by the idea of using a force-controlled gripper as a tactile sensor. This method enhances policy robustness without additional data collection and hardware involvement, while providing a noise-free binary gripper state observation for the policy and thus facilitating pure simulation learning to unleash the power of simulation. Experimental results across three real-world grasp-based tasks demonstrate the necessity, effectiveness, and efficiency of our approach.

comment: 8 pages, 5 figures, submitted to IROS 2025, project page: https://yifei-y.github.io/project-pages/Pseudo-Tactile-Feedback/

☆ Trajectory Planning for Automated Driving using Target Funnels

Benjamin Bogenberger, Johannes Bürger, Vladislav Nenchev

Self-driving vehicles rely on sensory input to monitor their surroundings and continuously adapt to the most likely future road course. Predictive trajectory planning is based on snapshots of the (uncertain) road course as a key input. Under noisy perception data, estimates of the road course can vary significantly, leading to indecisive and erratic steering behavior. To overcome this issue, this paper introduces a predictive trajectory planning algorithm with a novel objective function: instead of targeting a single reference trajectory based on the most likely road course, tracking a series of target reference sets, called a target funnel, is considered. The proposed planning algorithm integrates probabilistic information about the road course, and thus implicitly considers regular updates to road perception. Our solution is assessed in a case study using real driving data collected from a prototype vehicle. The results demonstrate that the algorithm maintains tracking accuracy and substantially reduces undesirable steering commands in the presence of noisy road perception, achieving a 56% reduction in input costs compared to a certainty equivalent formulation.

comment: accepted to European Control Conference 2025 (ECC25)

☆ Towards a cognitive architecture to enable natural language interaction in co-constructive task learning

Manuel Scheibl, Birte Richter, Alissa Müller, Michael Beetz, Britta Wrede

This research addresses the question, which characteristics a cognitive architecture must have to leverage the benefits of natural language in Co-Constructive Task Learning (CCTL). To provide context, we first discuss Interactive Task Learning (ITL), the mechanisms of the human memory system, and the significance of natural language and multi-modality. Next, we examine the current state of cognitive architectures, analyzing their capabilities to inform a concept of CCTL grounded in multiple sources. We then integrate insights from various research domains to develop a unified framework. Finally, we conclude by identifying the remaining challenges and requirements necessary to achieve CCTL in Human-Robot Interaction (HRI).

comment: 8 pages, 5 figures, submitted to: IEEE RO-MAN 2025

☆ Towards Benchmarking and Assessing the Safety and Robustness of Autonomous Driving on Safety-critical Scenarios

Jingzheng Li, Xianglong Liu, Shikui Wei, Zhijun Chen, Bing Li, Qing Guo, Xianqi Yang, Yanjun Pu, Jiakai Wang

Autonomous driving has made significant progress in both academia and industry, including performance improvements in perception task and the development of end-to-end autonomous driving systems. However, the safety and robustness assessment of autonomous driving has not received sufficient attention. Current evaluations of autonomous driving are typically conducted in natural driving scenarios. However, many accidents often occur in edge cases, also known as safety-critical scenarios. These safety-critical scenarios are difficult to collect, and there is currently no clear definition of what constitutes a safety-critical scenario. In this work, we explore the safety and robustness of autonomous driving in safety-critical scenarios. First, we provide a definition of safety-critical scenarios, including static traffic scenarios such as adversarial attack scenarios and natural distribution shifts, as well as dynamic traffic scenarios such as accident scenarios. Then, we develop an autonomous driving safety testing platform to comprehensively evaluate autonomous driving systems, encompassing not only the assessment of perception modules but also system-level evaluations. Our work systematically constructs a safety verification process for autonomous driving, providing technical support for the industry to establish standardized test framework and reduce risks in real-world road deployment.

☆ A Survey of Reinforcement Learning-Based Motion Planning for Autonomous Driving: Lessons Learned from a Driving Task Perspective

Zhuoren Li, Guizhe Jin, Ran Yu, Zhiwen Chen, Nan Li, Wei Han, Lu Xiong, Bo Leng, Jia Hu, Ilya Kolmanovsky, Dimitar Filev

Reinforcement learning (RL), with its ability to explore and optimize policies in complex, dynamic decision-making tasks, has emerged as a promising approach to addressing motion planning (MoP) challenges in autonomous driving (AD). Despite rapid advancements in RL and AD, a systematic description and interpretation of the RL design process tailored to diverse driving tasks remains underdeveloped. This survey provides a comprehensive review of RL-based MoP for AD, focusing on lessons from task-specific perspectives. We first outline the fundamentals of RL methodologies, and then survey their applications in MoP, analyzing scenario-specific features and task requirements to shed light on their influence on RL design choices. Building on this analysis, we summarize key design experiences, extract insights from various driving task applications, and provide guidance for future implementations. Additionally, we examine the frontier challenges in RL-based MoP, review recent efforts to addresse these challenges, and propose strategies for overcoming unresolved issues.

comment: 21 pages, 5 figures

♻ ☆ Robust Nonprehensile Object Transportation with Uncertain Inertial Parameters

Adam Heins, Angela P. Schoellig

We consider the nonprehensile object transportation task known as the waiter's problem - in which a robot must move an object on a tray from one location to another - when the transported object has uncertain inertial parameters. In contrast to existing approaches that completely ignore uncertainty in the inertia matrix or which only consider small parameter errors, we are interested in pushing the limits of the amount of inertial parameter uncertainty that can be handled. We first show how constraints that are robust to inertial parameter uncertainty can be incorporated into an optimization-based motion planning framework to transport objects while moving quickly. Next, we develop necessary conditions for the inertial parameters to be realizable on a bounding shape based on moment relaxations, allowing us to verify whether a trajectory will violate the constraints for any realizable inertial parameters. Finally, we demonstrate our approach on a mobile manipulator in simulations and real hardware experiments: our proposed robust constraints consistently successfully transport a 56 cm tall object with substantial inertial parameter uncertainty in the real world, while the baseline approaches drop the object while transporting it.

comment: 8 pages, 7 figures. Published in IEEE Robotics and Automation Letters

♻ ☆ CALMM-Drive: Confidence-Aware Autonomous Driving with Large Multimodal Model

Ruoyu Yao, Yubin Wang, Haichao Liu, Rui Yang, Zengqi Peng, Lei Zhu, Jun Ma

Decision-making and motion planning constitute critical components for ensuring the safety and efficiency of autonomous vehicles (AVs). Existing methodologies typically adopt two paradigms: decision then planning or generation then scoring. However, the former architecture often suffers from decision-planning misalignment that incurs risky situations. Meanwhile, the latter struggles to balance short-term operational metrics (e.g., immediate motion smoothness) with long-term tactical goals (e.g., route efficiency), resulting in myopic or overly conservative behaviors. To address these issues, we introduce CALMM-Drive, a novel Confidence-Aware Large Multimodal Model (LMM) empowered Autonomous Driving framework. Our approach integrates driving task-oriented Chain-of-Thought (CoT) reasoning coupled with Top-K confidence elicitation, which facilitates high-level reasoning to generate multiple candidate decisions with their confidence levels. Furthermore, we propose a novel planning module that integrates a diffusion model for trajectory generation and a hierarchical refinement process to find the optimal trajectory. This framework enables the selection over trajectory candidates accounting for both low-level solution quality and high-level tactical confidence, which avoids the risks within one-shot decisions and overcomes the limitations in short-sighted scoring mechanisms. Comprehensive evaluations in nuPlan closed-loop simulation environments demonstrate the competitive performance of CALMM-Drive across both common and long-tail benchmarks, showcasing a significant advancement in the integration of uncertainty in LMM-empowered AVs. The code will be released upon acceptance.

comment: 14 pages, 7 figures

♻ ☆ Tactile Ergodic Coverage on Curved Surfaces

Cem Bilaloglu, Tobias Löw, Sylvain Calinon

In this article, we present a feedback control method for tactile coverage tasks, such as cleaning or surface inspection. These tasks are challenging to plan due to complex continuous physical interactions. In these tasks, the coverage target and progress can be easily measured using a camera and encoded in a point cloud. We propose an ergodic coverage method that operates directly on point clouds, guiding the robot to spend more time on regions requiring more coverage. For robot control and contact behavior, we use geometric algebra to formulate a task-space impedance controller that tracks a line while simultaneously exerting a desired force along that line. We evaluate the performance of our method in kinematic simulations and demonstrate its applicability in real-world experiments on kitchenware. Our source codes, experimental data, and videos are available as open access at https://sites.google.com/view/tactile-ergodic-control/

♻ ☆ Fast and Accurate Task Planning using Neuro-Symbolic Language Models and Multi-level Goal Decomposition

Minseo Kwon, Yaesol Kim, Young J. Kim

In robotic task planning, symbolic planners using rule-based representations like PDDL are effective but struggle with long-sequential tasks in complicated environments due to exponentially increasing search space. Meanwhile, LLM-based approaches, which are grounded in artificial neural networks, offer faster inference and commonsense reasoning but suffer from lower success rates. To address the limitations of the current symbolic (slow speed) or LLM-based approaches (low accuracy), we propose a novel neuro-symbolic task planner that decomposes complex tasks into subgoals using LLM and carries out task planning for each subgoal using either symbolic or MCTS-based LLM planners, depending on the subgoal complexity. This decomposition reduces planning time and improves success rates by narrowing the search space and enabling LLMs to focus on more manageable tasks. Our method significantly reduces planning time while maintaining high success rates across task planning domains, as well as real-world and simulated robotics environments. More details are available at http://graphics.ewha.ac.kr/LLMTAMP/.

♻ ☆ Fast Online Learning of CLiFF-maps in Changing Environments ICRA

Yufei Zhu, Andrey Rudenko, Luigi Palmieri, Lukas Heuer, Achim J. Lilienthal, Martin Magnusson

Maps of dynamics are effective representations of motion patterns learned from prior observations, with recent research demonstrating their ability to enhance various downstream tasks such as human-aware robot navigation, long-term human motion prediction, and robot localization. Current advancements have primarily concentrated on methods for learning maps of human flow in environments where the flow is static, i.e., not assumed to change over time. In this paper we propose an online update method of the CLiFF-map (an advanced map of dynamics type that models motion patterns as velocity and orientation mixtures) to actively detect and adapt to human flow changes. As new observations are collected, our goal is to update a CLiFF-map to effectively and accurately integrate them, while retaining relevant historic motion patterns. The proposed online update method maintains a probabilistic representation in each observed location, updating parameters by continuously tracking sufficient statistics. In experiments using both synthetic and real-world datasets, we show that our method is able to maintain accurate representations of human motion dynamics, contributing to high performance flow-compliant planning downstream tasks, while being orders of magnitude faster than the comparable baselines.

comment: Accepted to the 2025 IEEE International Conference on Robotics and Automation (ICRA)

♻ ☆ Grasping a Handful: Sequential Multi-Object Dexterous Grasp Generation

Haofei Lu, Yifei Dong, Zehang Weng, Jens Lundell, Danica Kragic

We introduce the sequential multi-object robotic grasp sampling algorithm SeqGrasp that can robustly synthesize stable grasps on diverse objects using the robotic hand's partial Degrees of Freedom (DoF). We use SeqGrasp to construct the large-scale Allegro Hand sequential grasping dataset SeqDataset and use it for training the diffusion-based sequential grasp generator SeqDiffuser. We experimentally evaluate SeqGrasp and SeqDiffuser against the state-of-the-art non-sequential multi-object grasp generation method MultiGrasp in simulation and on a real robot. The experimental results demonstrate that SeqGrasp and SeqDiffuser reach an 8.71%-43.33% higher grasp success rate than MultiGrasp. Furthermore, SeqDiffuser is approximately 1000 times faster at generating grasps than SeqGrasp and MultiGrasp.

comment: 8 pages, 7 figures

♻ ☆ Dynamic High-Order Control Barrier Functions with Diffuser for Safety-Critical Trajectory Planning at Signal-Free Intersections

Di Chen, Ruiguo Zhong, Kehua Chen, Zhiwei Shang, Meixin Zhu, Edward Chung

Planning safe and efficient trajectories through signal-free intersections presents significant challenges for autonomous vehicles (AVs), particularly in dynamic, multi-task environments with unpredictable interactions and an increased possibility of conflicts. This study aims to address these challenges by developing a unified, robust, adaptive framework to ensure safety and efficiency across three distinct intersection movements: left-turn, right-turn, and straight-ahead. Existing methods often struggle to reliably ensure safety and effectively learn multi-task behaviors from demonstrations in such environments. This study proposes a safety-critical planning method that integrates Dynamic High-Order Control Barrier Functions (DHOCBF) with a diffusion-based model, called Dynamic Safety-Critical Diffuser (DSC-Diffuser). The DSC-Diffuser leverages task-guided planning to enhance efficiency, allowing the simultaneous learning of multiple driving tasks from real-world expert demonstrations. Moreover, the incorporation of goal-oriented constraints significantly reduces displacement errors, ensuring precise trajectory execution. To further ensure driving safety in dynamic environments, the proposed DHOCBF framework dynamically adjusts to account for the movements of surrounding vehicles, offering enhanced adaptability and reduce the conservatism compared to traditional control barrier functions. Validity evaluations of DHOCBF, conducted through numerical simulations, demonstrate its robustness in adapting to variations in obstacle velocities, sizes, uncertainties, and locations, effectively maintaining driving safety across a wide range of complex and uncertain scenarios. Comprehensive performance evaluations demonstrate that DSC-Diffuser generates realistic, stable, and generalizable policies, providing flexibility and reliable safety assurance in complex multi-task driving scenarios.

comment: 11 figures, 5 tables, 15 pages

♻ ☆ Mitigating Covariate Shift in Imitation Learning for Autonomous Vehicles Using Latent Space Generative World Models ICRA 2025

Alexander Popov, Alperen Degirmenci, David Wehr, Shashank Hegde, Ryan Oldja, Alexey Kamenev, Bertrand Douillard, David Nistér, Urs Muller, Ruchi Bhargava, Stan Birchfield, Nikolai Smolyanskiy

We propose the use of latent space generative world models to address the covariate shift problem in autonomous driving. A world model is a neural network capable of predicting an agent's next state given past states and actions. By leveraging a world model during training, the driving policy effectively mitigates covariate shift without requiring an excessive amount of training data. During end-to-end training, our policy learns how to recover from errors by aligning with states observed in human demonstrations, so that at runtime it can recover from perturbations outside the training distribution. Additionally, we introduce a novel transformer-based perception encoder that employs multi-view cross-attention and a learned scene query. We present qualitative and quantitative results, demonstrating significant improvements upon prior state of the art in closed-loop testing in the CARLA simulator, as well as showing the ability to handle perturbations in both CARLA and NVIDIA's DRIVE Sim.

comment: 8 pages, 6 figures, updated in March 2025, original published in September 2024, for ICRA 2025 submission, for associated video file, see https://youtu.be/7m3bXzlVQvU

♻ ☆ Beyond Omakase: Designing Shared Control for Navigation Robots with Blind People

Rie Kamikubo, Seita Kayukawa, Yuka Kaniwa, Allan Wang, Hernisa Kacorri, Hironobu Takagi, Chieko Asakawa

Autonomous navigation robots can increase the independence of blind people but often limit user control, following what is called in Japanese an "omakase" approach where decisions are left to the robot. This research investigates ways to enhance user control in social robot navigation, based on two studies conducted with blind participants. The first study, involving structured interviews (N=14), identified crowded spaces as key areas with significant social challenges. The second study (N=13) explored navigation tasks with an autonomous robot in these environments and identified design strategies across different modes of autonomy. Participants preferred an active role, termed the "boss" mode, where they managed crowd interactions, while the "monitor" mode helped them assess the environment, negotiate movements, and interact with the robot. These findings highlight the importance of shared control and user involvement for blind users, offering valuable insights for designing future social navigation robots.

comment: Preprint, ACM CHI Conference on Human Factors in Computing Systems (CHI 2025)

♻ ☆ Scalable Multi-modal Model Predictive Control via Duality-based Interaction Predictions

Hansung Kim, Siddharth H. Nair, Francesco Borrelli

We propose a hierarchical architecture designed for scalable real-time Model Predictive Control (MPC) in complex, multi-modal traffic scenarios. This architecture comprises two key components: 1) RAID-Net, a novel attention-based Recurrent Neural Network that predicts relevant interactions along the MPC prediction horizon between the autonomous vehicle and the surrounding vehicles using Lagrangian duality, and 2) a reduced Stochastic MPC problem that eliminates irrelevant collision avoidance constraints, enhancing computational efficiency. Our approach is demonstrated in a simulated traffic intersection with interactive surrounding vehicles, showcasing a 12x speed-up in solving the motion planning problem. A video demonstrating the proposed architecture in multiple complex traffic scenarios can be found here: https://youtu.be/-pRiOnPb9_c. GitHub: https://github.com/MPC-Berkeley/hmpc_raidnet

comment: Accepted at IEEE Intelligent Vehicles Symposium 2024

♻ ☆ Joint Moment Estimation for Hip Exoskeleton Control: A Generalized Moment Feature Generation Method

Yuanwen Zhang, Jingfeng Xiong, Haolan Xian, Chuheng Chen, Xinxing Chen, Chenglong Fu, Yuquan Leng

Hip joint moments during walking are the key foundation for hip exoskeleton assistance control. Most recent studies have shown estimating hip joint moments instantaneously offers a lot of advantages compared to generating assistive torque profiles based on gait estimation, such as simple sensor requirements and adaptability to variable walking speeds. However, existing joint moment estimation methods still suffer from a lack of personalization, leading to estimation accuracy degradation for new users. To address the challenges, this paper proposes a hip joint moment estimation method based on generalized moment features (GMF). A GMF generator is constructed to learn GMF of the joint moment which is invariant to individual variations while remaining decodable into joint moments through a dedicated decoder. Utilizing this well-featured representation, a GRU-based neural network is used to predict GMF with joint kinematics data, which can easily be acquired by hip exoskeleton encoders. The proposed estimation method achieves a root mean square error of 0.1180 Nm/kg under 28 walking speed conditions on a treadmill dataset, improved by 6.5% compared to the model without body parameter fusion, and by 8.3% for the conventional fusion model with body parameter. Furthermore, the proposed method was employed on a hip exoskeleton with only encoder sensors and achieved an average 20.5% metabolic reduction (p<0.01) for users compared to assist-off condition in level-ground walking.

comment: 13 pages, 10 figures, Submitted to Biomimetic Intelligence and Robotics

♻ ☆ Efficiently Generating Expressive Quadruped Behaviors via Language-Guided Preference Learning

Jaden Clark, Joey Hejna, Dorsa Sadigh

Expressive robotic behavior is essential for the widespread acceptance of robots in social environments. Recent advancements in learned legged locomotion controllers have enabled more dynamic and versatile robot behaviors. However, determining the optimal behavior for interactions with different users across varied scenarios remains a challenge. Current methods either rely on natural language input, which is efficient but low-resolution, or learn from human preferences, which, although high-resolution, is sample inefficient. This paper introduces a novel approach that leverages priors generated by pre-trained LLMs alongside the precision of preference learning. Our method, termed Language-Guided Preference Learning (LGPL), uses LLMs to generate initial behavior samples, which are then refined through preference-based feedback to learn behaviors that closely align with human expectations. Our core insight is that LLMs can guide the sampling process for preference learning, leading to a substantial improvement in sample efficiency. We demonstrate that LGPL can quickly learn accurate and expressive behaviors with as few as four queries, outperforming both purely language-parameterized models and traditional preference learning approaches. Website with videos: https://lgpl-gaits.github.io/

comment: 8 pages 5 figures

☆ Policy Gradient for LQR with Domain Randomization

Tesshu Fujinami, Bruce D. Lee, Nikolai Matni, George J. Pappas

Domain randomization (DR) enables sim-to-real transfer by training controllers on a distribution of simulated environments, with the goal of achieving robust performance in the real world. Although DR is widely used in practice and is often solved using simple policy gradient (PG) methods, understanding of its theoretical guarantees remains limited. Toward addressing this gap, we provide the first convergence analysis of PG methods for domain-randomized linear quadratic regulation (LQR). We show that PG converges globally to the minimizer of a finite-sample approximation of the DR objective under suitable bounds on the heterogeneity of the sampled systems. We also quantify the sample-complexity associated with achieving a small performance gap between the sample-average and population-level objectives. Additionally, we propose and analyze a discount-factor annealing algorithm that obviates the need for an initial jointly stabilizing controller, which may be challenging to find. Empirical results support our theoretical findings and highlight promising directions for future work, including risk-sensitive DR formulations and stochastic PG algorithms.

☆ Coordinating Distributed Energy Resources with Nodal Pricing in Distribution Networks: a Game-Theoretic Approach

Eli Brock, Jingqi Li, Javad Lavaei, Somayeh Sojoudi

We propose a real-time nodal pricing mechanism for cost minimization and voltage control in a distribution network with autonomous distributed energy resources and analyze the resulting market using stochastic game theory. Unlike existing methods, the proposed pricing scheme does not require device-aware centralized coordination or communication between prosumers. By developing new sufficient conditions under which a stochastic game is a Markov potential game, we show that the problem of computing an equilibrium for the proposed model is equivalent to solving a single-agent Markov Decision Process. These new conditions are general and may apply to other applications. We compute the equilibrium for an IEEE test system to empirically demonstrate the effectiveness of the pricing policy.

☆ Analysis of the French system imbalance paving the way for a novel operating reserve sizing approach

Jonathan Dumas, Sébastien Finet, Nathalie Grisey, Ibtissam Hamdane, Paul Plessiez

This paper examines the relationship between system imbalance and several explanatory variables within the French electricity system. The factors considered include lagged imbalance values, observations of renewable energy sources (RES) generation and consumption, and forecasts for RES generation and consumption. The study analyzes the distribution of system imbalance in relation to these variables. Additionally, an HGBR machine-learning model is employed to assess the predictability of imbalances and the explanatory power of the input variables studied. The results indicate no clear correlation between RES generation or consumption and the observed imbalances. However, it is possible to predict the imbalance adequately using forecasts available a few hours before real-time, along with the lagged values of the imbalance. Predicting the imbalance a day in advance proves to be complex with the variables examined; however, the extreme quantiles of the imbalance used for reserve sizing and contracting can be predicted with sufficient accuracy.

comment: Paper accepted to be presented at the EEM 2025 conference

☆ Disturbance-adaptive Model Predictive Control for Bounded Average Constraint Violations

Jicheng Shi, Colin N. Jones

This paper considers stochastic linear time-invariant systems subject to constraints on the average number of state-constraint violations over time without knowing the disturbance distribution. We present a novel disturbance-adaptive model predictive control (DAD-MPC) framework, which adjusts the disturbance model based on measured constraint violations. Using a robust invariance method, DAD-MPC ensures recursive feasibility and guarantees asymptotic or robust bounds on average constraint violations. Additionally, the bounds hold even with an inaccurate disturbance model, which allows for data-driven disturbance quantification methods to be used, such as conformal prediction. Simulation results demonstrate that the proposed approach outperforms state-of-the-art methods while satisfying average violation constraints.

☆ A system level approach to generalised feedback Nash equilibrium seeking in partially-observed games

Otacilio B. L. Neto, Michela Mulas, Francesco Corona

This work proposes an algorithm for seeking generalised feedback Nash equilibria (GFNE) in noncooperative dynamic games. The focus is on cyber-physical systems with dynamics which are linear, stochastic, potentially unstable, and partially observed. We employ System Level Synthesis (SLS) to reformulate the problem as the search for an equilibrium profile of closed-loop responses to noise, which can then be used to reconstruct a stabilising output-feedback policy. Under this setup, we leverage monotone operator theory to design a GFNE-seeking algorithm capable to enforce closed-loop stability, operational constraints, and communication constraints onto the control policies. This algorithm is amenable to numerical implementation and we provide conditions for its convergence. We demonstrate our approach in a simulated experiment on the noncooperative stabilisation of a decentralised power-grid.

☆ Quantifying Grid-Forming Behavior: Bridging Device-level Dynamics and System-Level Stability

Kehao Zhuang, Huanhai Xin, Verena Häberle, Xiuqiang He, Linbin Huang, Florian Dörfler

Grid-Forming (GFM) technology is considered a promising solution to build power electronics-dominated power systems. However, the impact of GFM converters on the system stability is still unquantified, creating a gap between the system- and device-level perspectives. To fill this gap, at the device-level, we propose a Forming Index to quantify a converter's response to grid voltage variations, providing a characterization of its GFM behavior. At the system-level, a quantitative notion of System Strength is introduced to capture the fundamental requirements for power system formation. Finally, we establish the alignment between device- and system-level metrics by demonstrating that GFM converters provably enhance system strength.

☆ Robust Feedback Optimization with Model Uncertainty: A Regularization Approach

Winnie Chan, Zhiyu He, Keith Moffat, Saverio Bolognani, Michael Muehlebach, Florian Dörfler

Feedback optimization optimizes the steady state of a dynamical system by implementing optimization iterations in closed loop with the plant. It relies on online measurements and limited model information, namely, the input-output sensitivity. In practice, various issues including inaccurate modeling, lack of observation, or changing conditions can lead to sensitivity mismatches, causing closed-loop sub-optimality or even instability. To handle such uncertainties, we pursue robust feedback optimization, where we optimize the closed-loop performance against all possible sensitivities lying in specific uncertainty sets. We provide tractable reformulations for the corresponding min-max problems via regularizations and characterize the online closed-loop performance through the tracking error in case of time-varying optimal solutions. Simulations on a distribution grid illustrate the effectiveness of our robust feedback optimization controller in addressing sensitivity mismatches in a non-stationary environment.

☆ Data-Driven Distributed Output Synchronization of Heterogeneous Discrete-Time Multi-Agent Systems

Giulio Fattore, Maria Elena Valcher

In this paper, we assume that an autonomous exosystem generates a reference output, and we consider the problem of designing a distributed data-driven control law for a family of discrete-time heterogeneous LTI agents, connected through a directed graph, in order to synchronize the agents' outputs to the reference one. The agents of the network are split into two categories: leaders, with direct access to the exosystem output, and followers, that only receive information from their neighbors. All agents aim to achieve output synchronization by means of a state feedback that makes use of their own states as well as of an estimate of the exogenous system state, provided by an internal state observer. Such observer has a different structure for leaders and followers. Necessary and sufficient conditions for the existence of a solution are first derived in the model-based set-up and then in a data-driven context. An example illustrates both the implementation procedure and the performance of the proposed approach.

comment: Extended version of the conference paper submitted to 64th IEEE Conference on Decision and Control

☆ Application of Battery Storage to Switching Predictive Control of Power Distribution Systems Including Road Heating

Chiaki Kojima, Yuya Muto, Hikaru Akutsu, Rinnosuke Shima, Yoshihiko Susuki

A road heating system is an electrical device which promotes snow melting by burying a heating cable as a thermal source underground. When integrating road heating into the power distribution system, we need to optimize the flow of electric power by appropriately integrating distributed power sources and conventional power distribution equipment. In this paper, we extend the power distribution system considered in the authors' previous study to the case where battery storage is installed. As a main result, we propose a predictive switching control that achieves the reduction of distribution loss, attenuation of voltage fluctuation, and efficient snow melting, simultaneously. We verify the effectiveness of the application of battery storage through numerical simulation.

comment: 13 pages, 14 figures

☆ Initial State Privacy of Nonlinear Systems on Riemannian Manifolds

Le Liu, Yu Kawano, Antai Xie, Ming Cao

In this paper, we investigate initial state privacy protection for discrete-time nonlinear closed systems. By capturing Riemannian geometric structures inherent in such privacy challenges, we refine the concept of differential privacy through the introduction of an initial state adjacency set based on Riemannian distances. A new differential privacy condition is formulated using incremental output boundedness, enabling the design of time-varying Laplacian noise to achieve specified privacy guarantees. The proposed framework extends beyond initial state protection to also cover system parameter privacy, which is demonstrated as a special application.

☆ Distributed AC Optimal Power Flow: A Scalable Solution for Large-Scale Problems

Xinliang Dai, Yuning Jiang, Yi Guo, Colin N. Jones, Moritz Diehl, Veit Hagenmeyer

This paper introduces a novel distributed optimization framework for large-scale AC Optimal Power Flow (OPF) problems, offering both theoretical convergence guarantees and rapid convergence in practice. By integrating smoothing techniques and the Schur complement, the proposed approach addresses the scalability challenges and reduces communication overhead in distributed AC OPF. Additionally, optimal network decomposition enables efficient parallel processing under the single program multiple data (SPMD) paradigm. Extensive simulations on large-scale benchmarks across various operating scenarios indicate that the proposed framework outperforms the state-of-the-art centralized solver IPOPT on modest hardware. This paves the way for more scalable and efficient distributed optimization in future power system applications.

☆ Unraveling tensor structures in correct-by-design controller synthesis

Ruohan Wang, Zhiyong Sun, Sofie Haesaert

Formal safety guarantees on the synthesis of controllers for stochastic systems can be obtained using correct-by-design approaches. These approaches often use abstractions as finite-state Markov Decision Processes. As the state space of these MDPs grows, the curse of dimensionality makes the computational and memory cost of the probabilistic guarantees, quantified with dynamic programming, scale exponentially. In this work, we leverage decoupled dynamics and unravel, via dynamic programming operations, a tree structure in the Canonical Polyadic Decomposition (CPD) of the value functions. For discrete-time stochastic systems with syntactically co-safe linear temporal logic (scLTL) specifications, we provide provable probabilistic safety guarantees and significantly alleviate the computational burden. We provide an initial validation of the theoretical results on several typical case studies and showcase that the uncovered tree structure enables efficient reductions in the computational burden.

☆ A robot-assisted pipeline to rapidly scan 1.7 million historical aerial photographs

Sheila Masson, Alan Potts, Allan Williams, Steve Berggreen, Kevin McLaren, Sam Martin, Eugenio Noda, Nicklas Nordfors, Nic Ruecroft, Hannah Druckenmiller, Solomon Hsiang, Andreas Madestam, Anna Tompsett

During the 20th Century, aerial surveys captured hundreds of millions of high-resolution photographs of the earth's surface. These images, the precursors to modern satellite imagery, represent an extraordinary visual record of the environmental and social upheavals of the 20th Century. However, most of these images currently languish in physical archives where retrieval is difficult and costly. Digitization could revolutionize access, but manual scanning is slow and expensive. Here, we describe and validate a novel robot-assisted pipeline that increases worker productivity in scanning 30-fold, applied at scale to digitize an archive of 1.7 million historical aerial photographs from 65 countries.

☆ An ANN-Enhanced Approach for Flatness-Based Constrained Control of Nonlinear Systems

Huu-Thinh Do, Ionela Prodan, Florin Stoican

Neural networks have proven practical for a synergistic combination of advanced control techniques. This work analyzes the implementation of rectified linear unit neural networks to achieve constrained control in differentially flat systems. Specifically, the class of flat systems enjoys the benefit of feedback linearizability, i.e., the systems can be linearized by means of a proper variable transformation. However, the price for linearizing the dynamics is that the constraint descriptions are distorted geometrically. Our results show that, by using neural networks, these constraints can be represented as a union of polytopes, enabling the use of mixed-integer programming tools to guarantee constraint satisfaction. We further analyze the integration of the characterization into efficient settings such as control Lyapunov function-based and model predictive control (MPC). Interestingly, this description also allows us to explicitly compute the solution of the MPC problem for the nonlinear system. Several examples are provided to illustrate the effectiveness of our framework.

☆ Consensus on Open Multi-Agent Systems Over Graphs Sampled from Graphons

Renato Vizuete, Julien M. Hendrickx

We show how graphons can be used to model and analyze open multi-agent systems, which are multi-agent systems subject to arrivals and departures, in the specific case of linear consensus. First, we analyze the case of replacements, where under the assumption of a deterministic interval between two replacements, we derive an upper bound for the disagreement in expectation. Then, we study the case of arrivals and departures, where we define a process for the evolution of the number of agents that guarantees a minimum and a maximum number of agents. Next, we derive an upper bound for the disagreement in expectation, and we establish a link with the spectrum of the expected graph used to generate the graph topologies. Finally, for stochastic block model (SBM) graphons, we prove that the computation of the spectrum of the expected graph can be performed based on a matrix whose dimension depends only on the graphon and it is independent of the number of agents.

comment: 8 pages, 1 figure

☆ Two-wheel-driven Electric Superbike Powertrain Optimization

Adelmo Niccolai, Maurizio Clemente, Theo Hofman, Niccolò Baldanzini

In this paper, we propose an optimization framework for the powertrain design of a two-wheel-driven electric superbike, minimizing energy consumption. Specifically, we jointly optimize the force distribution between the wheels with the gear ratio, and rear motor and battery sizing while explicitly considering vehicle dynamics and performance constraints. First, we present an energy consumption model of the vehicle, including a scalable model of the electric machine based on data from the industry, accounting for iron, copper, and mechanical losses. Then, we analyze the propulsive blending strategy to distribute the required power to the wheels while considering adherence limits. Finally, we demonstrate the effectiveness of our approach by analyzing the design of a superbike, based on regulatory driving cycles and a custom high-performance circuit by comparing the force distribution approaches. The results underline the significance of joint optimization of powertrain components and propulsive bias, achieving a reduction of up to 22.36% in energy consumption for the Sport high-performance driving cycle.

comment: 6 pages, 3 figures, 3 tables

☆ Distributionally Robust Model Order Reduction for Linear Systems

Le Liu, Yu Kawano, Yangming Dou, Ming Cao

In this paper, we investigate distributionally robust model order reduction for linear, discrete-time, time-invariant systems. The external input is assumed to follow an uncertain distribution within a Wasserstein ambiguity set. We begin by considering the case where the distribution is certain and formulate an optimization problem to obtain the reduced model. When the distribution is uncertain, the interaction between the reduced-order model and the distribution is modeled by a Stackelberg game. To ensure solvability, we first introduce the Gelbrich distance and demonstrate that the Stackelberg game within a Wasserstein ambiguity set is equivalent to that within a Gelbrich ambiguity set. Then, we propose a nested optimization problem to solve the Stackelberg game. Furthermore, the nested optimization problem is relaxed into a nested convex optimization problem, ensuring computational feasibility. Finally, a simulation is presented to illustrate the effectiveness of the proposed method.

☆ Certified Approximate Reachability (CARe): Formal Error Bounds on Deep Learning of Reachable Sets

Prashant Solanki, Nikolaus Vertovec, Yannik Schnitzer, Jasper Van Beers, Coen de Visser, Alessandro Abate

Recent approaches to leveraging deep learning for computing reachable sets of continuous-time dynamical systems have gained popularity over traditional level-set methods, as they overcome the curse of dimensionality. However, as with level-set methods, considerable care needs to be taken in limiting approximation errors, particularly since no guarantees are provided during training on the accuracy of the learned reachable set. To address this limitation, we introduce an epsilon-approximate Hamilton-Jacobi Partial Differential Equation (HJ-PDE), which establishes a relationship between training loss and accuracy of the true reachable set. To formally certify this approximation, we leverage Satisfiability Modulo Theories (SMT) solvers to bound the residual error of the HJ-based loss function across the domain of interest. Leveraging Counter Example Guided Inductive Synthesis (CEGIS), we close the loop around learning and verification, by fine-tuning the neural network on counterexamples found by the SMT solver, thus improving the accuracy of the learned reachable set. To the best of our knowledge, Certified Approximate Reachability (CARe) is the first approach to provide soundness guarantees on learned reachable sets of continuous dynamical systems.

☆ Privacy Preservation for Statistical Input in Dynamical Systems

Le Liu, Yu Kawano, Ming Cao

This paper addresses the challenge of privacy preservation for statistical inputs in dynamical systems. Motivated by an autonomous building application, we formulate a privacy preservation problem for statistical inputs in linear time-invariant systems. What makes this problem widely applicable is that the inputs, rather than being assumed to be deterministic, follow a probability distribution, inherently embedding privacy-sensitive information that requires protection. This formulation also presents a technical challenge as conventional differential privacy mechanisms are not directly applicable. Through rigorous analysis, we develop strategy to achieve $(0, \delta)$ differential privacy through adding noise. Finally, the effectiveness of our methods is demonstrated by revisiting the autonomous building application.

☆ Surveying Uncertainty Representation: A Unified Model for Cyber-Physical Systems

Johannes Mäkelburg, Diego Perez-Palacin, Raffaela Mirandola, Maribel Acosta

Cyber-Physical Systems (CPS) operate in dynamic environments, leading to different types of uncertainty. This work provides a comprehensive review of uncertainty representations and categorizes them based on the dimensions used to represent uncertainty. Through this categorization, key gaps and limitations in existing approaches are identified. To address these issues, a Conceptual Model of Uncertainty Representations in CPS is introduced, integrating and extending existing models. Its applicability is demonstrated through examples from the automotive domain, showing its effectiveness in capturing and structuring uncertainty in real-world scenarios.

☆ Less is More: Contextual Sampling for Nonlinear Data-Enabled Predictive Control IROS 2025

Julius Beerwerth, Bassam Alrifaee

Data-enabled Predictive Control (DeePC) is a powerful data-driven approach for predictive control without requiring an explicit system model. However, its high computational cost limits its applicability to real-time robotic systems. For robotic applications such as motion planning and trajectory tracking, real-time control is crucial. Nonlinear DeePC either relies on large datasets or learning the nonlinearities to ensure predictive accuracy, leading to high computational complexity. This work introduces contextual sampling, a novel data selection strategy to handle nonlinearities for DeePC by dynamically selecting the most relevant data at each time step. By reducing the dataset size while preserving prediction accuracy, our method improves computational efficiency, of DeePC for real-time robotic applications. We validate our approach for autonomous vehicle motion planning. For a dataset size of 100 sub-trajectories, Contextual sampling DeePC reduces tracking error by 53.2 % compared to Leverage Score sampling. Additionally, Contextual sampling reduces max computation time by 87.2 % compared to using the full dataset of 491 sub-trajectories while achieving comparable tracking performance. These results highlight the potential of Contextual sampling to enable real-time, data-driven control for robotic systems.

comment: Submitted to IROS 2025 on March 1st

☆ Robust Suboptimal Local Basis Function Algorithms for Identification of Nonstationary FIR Systems in Impulsive Noise Environments

Maciej Niedźwiecki, Artur Gańcza, Wojciech Żuławiński, Agnieszka Wyłomańska

While local basis function (LBF) estimation algorithms, commonly used for identifying/tracking systems with time-varying parameters, demonstrate good performance under the assumption of normally distributed measurement noise, the estimation results may significantly deviate from satisfactory when the noise distribution is impulsive in nature, for example, corrupted by outliers. This paper introduces a computationally efficient method to make the LBF estimator robust, enhancing its resistance to impulsive noise. First, the choice of basis functions is optimized based on the knowledge of parameter variation statistics. Then, the parameter tracking algorithm is made robust using the sequential data trimming technique. Finally, it is demonstrated that the proposed algorithm can undergo online tuning through parallel estimation and leave-one-out cross-validation.

☆ Incremental capacity-based multi-feature fusion model for predicting state-of-health of lithium-ion batteries

Chenyu Jia, Guangshu Xia, Yuanhao Shi, Jianfang Jia, Jie Wen, Jianchao Zeng

Lithium-ion batteries have become an indispensable part of human industrial production and daily life. For the safe use, management and maintenance of lithium-ion batteries, the state of health (SOH) of lithium-ion batteries is an important indicator so that the SOH estimation is of significant practical value. In order to accurately predict SOH, this paper proposes a fusion prediction model which combines particle swarm optimization (PSO) algorithm, bi-directional long-short time memory network (BiLSTM) and adaptive boosting (AdaBoost) algorithm. In the proposed prediction model, indirect health indicators (HIs), which characterize battery degradation, are obtained with the help of incremental capacity analysis (ICA), and is fed into BiLSTM to extract time-series features, whose parameters are optimized by employing PSO algorithm. On this basis, the AdaBoost algorithm is applied to reduce the risk of overfitting the PSO-BiLSTM model. The study based on lithium-ion battery data from Center for Advanced Life Cycle Engineering (CALCE) shows that the PSO-BiLSTM-AdaBoost model has higher accuracy, better robustness, and generalization ability.

☆ Free Parametrization of L2-bounded State Space Models

Leonardo Massai, Giancarlo Ferrari-Trecate

Structured state-space models (SSMs) have emerged as a powerful architecture in machine learning and control, featuring stacked layers where each consists of a linear time-invariant (LTI) discrete-time system followed by a nonlinearity. While SSMs offer computational efficiency and excel in long-sequence predictions, their widespread adoption in applications like system identification and optimal control is hindered by the challenge of ensuring their stability and robustness properties. We introduce L2RU, a novel parametrization of SSMs that guarantees input-output stability and robustness by enforcing a prescribed L-bound for all parameter values. This design eliminates the need for complex constraints, allowing unconstrained optimization over L2RUs by using standard methods such as gradient descent. Leveraging tools from system theory and convex optimization, we derive a non-conservative parametrization of square discrete-time LTI systems with a specified L2-bound, forming the foundation of the L2RU architecture. Additionally, we enhance its performance with a bespoke initialization strategy optimized for long input sequences. Through a system identification task, we validate L2RU's superior performance, showcasing its potential in learning and control applications.

comment: 8 pages

☆ On the Analysis of Qualitative Nyquist Plots

Davide Tebaldi, Roberto Zanasi

A powerful tool in control and systems engineering is represented by Nyquist plots, for which a qualitative representation often gives a clearer visualization of the frequency response function that is typically not given by computer programs, especially if portions of the Nyquist plot extend to infinity. This letter addresses the graphical analysis of the frequency response function, with the objective of enhancing the procedure for the qualitative construction of Nyquist plots. Several results supported by analytical proofs are derived for what concerns the low and high frequency behavior, which enable to improve the qualitative construction of Nyquist plots in the vicinity of the initial and final points.

☆ Trajectory Planning for Automated Driving using Target Funnels

Benjamin Bogenberger, Johannes Bürger, Vladislav Nenchev

Self-driving vehicles rely on sensory input to monitor their surroundings and continuously adapt to the most likely future road course. Predictive trajectory planning is based on snapshots of the (uncertain) road course as a key input. Under noisy perception data, estimates of the road course can vary significantly, leading to indecisive and erratic steering behavior. To overcome this issue, this paper introduces a predictive trajectory planning algorithm with a novel objective function: instead of targeting a single reference trajectory based on the most likely road course, tracking a series of target reference sets, called a target funnel, is considered. The proposed planning algorithm integrates probabilistic information about the road course, and thus implicitly considers regular updates to road perception. Our solution is assessed in a case study using real driving data collected from a prototype vehicle. The results demonstrate that the algorithm maintains tracking accuracy and substantially reduces undesirable steering commands in the presence of noisy road perception, achieving a 56% reduction in input costs compared to a certainty equivalent formulation.

comment: accepted to European Control Conference 2025 (ECC25)

☆ On the Steady-State Distributionally Robust Kalman Filter

Minhyuk Jang, Astghik Hakobyan, Insoon Yang

State estimation in the presence of uncertain or data-driven noise distributions remains a critical challenge in control and robotics. Although the Kalman filter is the most popular choice, its performance degrades significantly when distributional mismatches occur, potentially leading to instability or divergence. To address this limitation, we introduce a novel steady-state distributionally robust (DR) Kalman filter that leverages Wasserstein ambiguity sets to explicitly account for uncertainties in both process and measurement noise distributions. Our filter achieves computational efficiency by requiring merely the offline solution of a single convex semidefinite program, which yields a constant DR Kalman gain for robust state estimation under distributional mismatches. Additionally, we derive explicit theoretical conditions on the ambiguity set radius that ensure the asymptotic convergence of the time-varying DR Kalman filter to the proposed steady-state solution. Numerical simulations demonstrate that our approach outperforms existing baseline filters in terms of robustness and accuracy across both Gaussian and non-Gaussian uncertainty scenarios, highlighting its significant potential for real-world control and estimation applications.

☆ Stability and Controllability of Revenue Systems via the Bode Approach

Yichuan Niu, Jianhui Chen

In online revenue systems, e.g. an advertising system, budget pacing plays a critical role in ensuring that the spend aligns with desired financial objectives. Pacing systems dynamically control the velocity of spending to balance auction intensity, traffic fluctuations, and other stochastic variables. Current industry practices rely heavily on trial-and-error approaches, often leading to inefficiencies and instability. This paper introduces a principled methodology rooted in Classical Control Theory to address these challenges. By modeling the pacing system as a linear time-invariant (LTI) proxy and leveraging compensator design techniques using Bode methodology, we derive a robust controller to minimize pacing errors and enhance stability. The proposed methodology is validated through simulation and tested by our in-house auction system, demonstrating superior performance in achieving precise budget allocation while maintaining resilience to traffic and auction dynamics. Our findings bridge the gap between traditional control theory and modern advertising systems in modeling, simulation, and validation, offering a scalable and systematic approach to budget pacing optimization.

☆ Optimizing Age of Information in Networks with Large and Small Updates

Zhuoyi Zhao, Vishrant Tripathi, Igor Kadota

Modern sensing and monitoring applications typically consist of sources transmitting updates of different sizes, ranging from a few bytes (position, temperature, etc.) to multiple megabytes (images, video frames, LIDAR point scans, etc.). Existing approaches to wireless scheduling for information freshness typically ignore this mix of large and small updates, leading to suboptimal performance. In this paper, we consider a single-hop wireless broadcast network with sources transmitting updates of different sizes to a base station over unreliable links. Some sources send large updates spanning many time slots while others send small updates spanning only a few time slots. Due to medium access constraints, only one source can transmit to the base station at any given time, thus requiring careful design of scheduling policies that takes the sizes of updates into account. First, we derive a lower bound on the achievable Age of Information (AoI) by any transmission scheduling policy. Second, we develop optimal randomized policies that consider both switching and no-switching during the transmission of large updates. Third, we introduce a novel Lyapunov function and associated analysis to propose an AoI-based Max-Weight policy that has provable constant factor optimality guarantees. Finally, we evaluate and compare the performance of our proposed scheduling policies through simulations, which show that our Max-Weight policy achieves near-optimal AoI performance.

comment: To appear in WiOpt 2025

☆ Remarks on the Polyak-Lojasiewicz inequality and the convergence of gradient systems

Arthur Castello B. de Oliveira, Leilei Cui, Eduardo D. Sontag

This work explores generalizations of the Polyak-Lojasiewicz inequality (PLI) and their implications for the convergence behavior of gradient flows in optimization problems. Motivated by the continuous-time linear quadratic regulator (CT-LQR) policy optimization problem -- where only a weaker version of the PLI is characterized in the literature -- this work shows that while weaker conditions are sufficient for global convergence to, and optimality of the set of critical points of the cost function, the "profile" of the gradient flow solution can change significantly depending on which "flavor" of inequality the cost satisfies. After a general theoretical analysis, we focus on fitting the CT-LQR policy optimization problem to the proposed framework, showing that, in fact, it can never satisfy a PLI in its strongest form. We follow up our analysis with a brief discussion on the difference between continuous- and discrete-time LQR policy optimization, and end the paper with some intuition on the extension of this framework to optimization problems with L1 regularization and solved through proximal gradient flows.

♻ ☆ Feasibility Evaluation of Quadratic Programs for Constrained Control

Panagiotis Rousseas, Dimitra Panagou

This paper presents a computationally-efficient method for evaluating the feasibility of Quadratic Programs (QPs) for online constrained control. Based on the duality principle, we first show that the feasibility of a QP can be determined by the solution of a properly-defined Linear Program (LP). Our analysis yields a LP that can be solved more efficiently compared to the original QP problem, and more importantly, is simpler in form and can be solved more efficiently compared to existing methods that assess feasibility via LPs. The computational efficiency of the proposed method compared to existing methods for feasibility evaluation is demonstrated in comparative case studies as well as a feasible-constraint selection problem, indicating its promise for online feasibility evaluation of optimization-based controllers.

comment: Submitted to CDC 2025

♻ ☆ Safety Filter Design for Articulated Frame Steering Vehicles In the Presence of Actuator Dynamics Using High-Order Control Barrier Functions

Naeim Ebrahimi Toulkani, Reza Ghabcheloo

Articulated Frame Steering (AFS) vehicles are widely used in heavy-duty industries, where they often operate near operators and laborers. Therefore, designing safe controllers for AFS vehicles is essential. In this paper, we develop a Quadratic Program (QP)-based safety filter that ensures feasibility for AFS vehicles with affine actuator dynamics. To achieve this, we first derive the general equations of motion for AFS vehicles, incorporating affine actuator dynamics. We then introduce a novel High-Order Control Barrier Function (HOCBF) candidate with equal relative degrees for both system controls. Finally, we design a Parametric Adaptive HOCBF (PACBF) and an always-feasible, QP-based safety filter. Numerical simulations of AFS vehicle kinematics demonstrate the effectiveness of our approach.

♻ ☆ Minimally Conservative Controlled-Invariant Set Synthesis Using Control Barrier Certificates

Naeim Ebrahimi Toulkani, Reza Ghabcheloo

Finding a controlled-invariant set for a system with state and control constraints is crucial for safety-critical applications. However, existing methods often produce overly conservative solutions. This paper presents a method for generating controlled-invariant (safe) sets for nonlinear polynomial control-affine systems using Control Barrier Certificates (CBCs). We formulate CBC conditions as Sum-of-Squares (SOS) constraints and solve them via an SOS Program (SOSP). First, we generalize existing SOSPs for CBC synthesis to handle environments with complex unsafe state representations. Then, we propose an iterative algorithm that progressively enlarges the safe set constructed by the synthesized CBCs by maximizing boundary expansion at each iteration. We theoretically prove that our method guarantees strict safe set expansion at every step. Finally, we validate our approach with numerical simulations in 2D and 3D for single-input and multi-input systems. Empirical results show that the safe set generated by our method covers in most part a larger portion of the state space compared to two state-of-the-art techniques.

♻ ☆ Understanding long-term energy use in off-grid solar home systems in sub-Saharan Africa

Rebecca Perriment, Vasco Mergulhao, Volkan Kumtepeli, Priti Parikh, Malcolm McCulloch, David Howey

Solar home systems provide low-cost electricity access for rural off-grid communities. As access to them increases, more long-term data becomes available on how these systems are used throughout their lifetime. This work analyses a dataset of 1,000 systems across sub-Saharan Africa. Dynamic time warping clustering was applied to the load demand data from the systems, identifying five distinct archetypal daily load profiles and their occurrence across the dataset. Temporal analysis reveals a general decline in daily energy consumption over time, with 77% of households reducing their usage compared to the start of ownership. On average, there is a 33% decrease in daily consumption by the end of the second year compared to the peak demand, which occurs on the 96th day. Combining the load demand analysis with payment data shows that this decrease in energy consumption is observed even in households that are not experiencing economic hardship, indicating there are reasons beyond financial constraints for decreasing energy use once energy access is obtained.

comment: Draft updates, including text and figure changes

♻ ☆ Dynamic High-Order Control Barrier Functions with Diffuser for Safety-Critical Trajectory Planning at Signal-Free Intersections

Di Chen, Ruiguo Zhong, Kehua Chen, Zhiwei Shang, Meixin Zhu, Edward Chung

Planning safe and efficient trajectories through signal-free intersections presents significant challenges for autonomous vehicles (AVs), particularly in dynamic, multi-task environments with unpredictable interactions and an increased possibility of conflicts. This study aims to address these challenges by developing a unified, robust, adaptive framework to ensure safety and efficiency across three distinct intersection movements: left-turn, right-turn, and straight-ahead. Existing methods often struggle to reliably ensure safety and effectively learn multi-task behaviors from demonstrations in such environments. This study proposes a safety-critical planning method that integrates Dynamic High-Order Control Barrier Functions (DHOCBF) with a diffusion-based model, called Dynamic Safety-Critical Diffuser (DSC-Diffuser). The DSC-Diffuser leverages task-guided planning to enhance efficiency, allowing the simultaneous learning of multiple driving tasks from real-world expert demonstrations. Moreover, the incorporation of goal-oriented constraints significantly reduces displacement errors, ensuring precise trajectory execution. To further ensure driving safety in dynamic environments, the proposed DHOCBF framework dynamically adjusts to account for the movements of surrounding vehicles, offering enhanced adaptability and reduce the conservatism compared to traditional control barrier functions. Validity evaluations of DHOCBF, conducted through numerical simulations, demonstrate its robustness in adapting to variations in obstacle velocities, sizes, uncertainties, and locations, effectively maintaining driving safety across a wide range of complex and uncertain scenarios. Comprehensive performance evaluations demonstrate that DSC-Diffuser generates realistic, stable, and generalizable policies, providing flexibility and reliable safety assurance in complex multi-task driving scenarios.

comment: 11 figures, 5 tables, 15 pages

♻ ☆ Mitigating Covariate Shift in Imitation Learning for Autonomous Vehicles Using Latent Space Generative World Models ICRA 2025

Alexander Popov, Alperen Degirmenci, David Wehr, Shashank Hegde, Ryan Oldja, Alexey Kamenev, Bertrand Douillard, David Nistér, Urs Muller, Ruchi Bhargava, Stan Birchfield, Nikolai Smolyanskiy

We propose the use of latent space generative world models to address the covariate shift problem in autonomous driving. A world model is a neural network capable of predicting an agent's next state given past states and actions. By leveraging a world model during training, the driving policy effectively mitigates covariate shift without requiring an excessive amount of training data. During end-to-end training, our policy learns how to recover from errors by aligning with states observed in human demonstrations, so that at runtime it can recover from perturbations outside the training distribution. Additionally, we introduce a novel transformer-based perception encoder that employs multi-view cross-attention and a learned scene query. We present qualitative and quantitative results, demonstrating significant improvements upon prior state of the art in closed-loop testing in the CARLA simulator, as well as showing the ability to handle perturbations in both CARLA and NVIDIA's DRIVE Sim.

comment: 8 pages, 6 figures, updated in March 2025, original published in September 2024, for ICRA 2025 submission, for associated video file, see https://youtu.be/7m3bXzlVQvU

♻ ☆ Stability results for MIMO LTI systems via Scaled Relative Graphs

Eder Baron-Prada, Adolfo Anta, Alberto Padoan, Florian Dörfler

This paper proposes a new approach for stability analysis of multi-input, multi-output (MIMO) feedback systems through Scaled Relative Graphs (SRGs). Unlike traditional methods, such as the Generalized Nyquist Criterion (GNC), which relies on a coupled analysis that requires the multiplication of models, our approach enables the evaluation of system stability in a decoupled fashion and provides an intuitive, visual representation of system behavior. Our results provide conditions for certifying the stability of feedback MIMO Linear Time-Invariant (LTI) systems.

comment: Submitted to CDC 2025

♻ ☆ Algorithmic analysis of systems with affine input and polynomial state

Lorenzo Clemente

The goal of this paper is to provide exact and terminating algorithms for the formal analysis of deterministic continuous-time control systems with affine input and polynomial state dynamics (in short, polynomial systems). We consider the following semantic properties: zeroness and equivalence, input independence, linearity, and analyticity. Our approach is based on Chen-Fliess series, which provide a unique representation of the dynamics of such systems via their formal generating series. Our starting point is Fliess' seminal work showing how the semantic properties above are mirrored by corresponding combinatorial properties on generating series. Next, we observe that the generating series of polynomial systems coincide with the class of shuffle-finite series, a nonlinear generalisation of Sch\"utzenberger's rational series which has recently been studied in the context of automata theory and enumerative combinatorics. We exploit and extend recent results in the algorithmic analysis of shuffle-finite series (such as zeroness, equivalence, and commutativity) to show that the semantic properties above can be decided exactly and in finite time for polynomial systems. Some of our analyses rely on a novel technical contribution, namely that shuffle-finite series are closed under support restrictions with commutative regular languages, a result of independent interest.

comment: technical report

♻ ☆ Scalable Multi-modal Model Predictive Control via Duality-based Interaction Predictions

Hansung Kim, Siddharth H. Nair, Francesco Borrelli

We propose a hierarchical architecture designed for scalable real-time Model Predictive Control (MPC) in complex, multi-modal traffic scenarios. This architecture comprises two key components: 1) RAID-Net, a novel attention-based Recurrent Neural Network that predicts relevant interactions along the MPC prediction horizon between the autonomous vehicle and the surrounding vehicles using Lagrangian duality, and 2) a reduced Stochastic MPC problem that eliminates irrelevant collision avoidance constraints, enhancing computational efficiency. Our approach is demonstrated in a simulated traffic intersection with interactive surrounding vehicles, showcasing a 12x speed-up in solving the motion planning problem. A video demonstrating the proposed architecture in multiple complex traffic scenarios can be found here: https://youtu.be/-pRiOnPb9_c. GitHub: https://github.com/MPC-Berkeley/hmpc_raidnet

comment: Accepted at IEEE Intelligent Vehicles Symposium 2024

♻ ☆ Cascade Reinforcement Learning with State Space Factorization for O-RAN-based Traffic Steering

Chuanneng Sun, Gueyoung Jung, Tuyen Xuan Tran, Dario Pompili

The Open Radio Access Network (O-RAN) architecture empowers intelligent and automated optimization of the RAN through applications deployed on the RAN Intelligent Controller (RIC) platform, enabling capabilities beyond what is achievable with traditional RAN solutions. Within this paradigm, Traffic Steering (TS) emerges as a pivotal RIC application that focuses on optimizing cell-level mobility settings in near-real-time, aiming to significantly improve network spectral efficiency. In this paper, we design a novel TS algorithm based on a Cascade Reinforcement Learning (CaRL) framework. We propose state space factorization and policy decomposition to reduce the need for large models and well-labeled datasets. For each sub-state space, an RL sub-policy will be trained to learn an optimized mapping onto the action space. To apply CaRL on new network regions, we propose a knowledge transfer approach to initialize a new sub-policy based on knowledge learned by the trained policies. To evaluate CaRL, we build a data-driven and scalable RIC digital twin (DT) that is modeled using important real-world data, including network configuration, user geo-distribution, and traffic demand, among others, from a tier-1 mobile operator in the US. We evaluate CaRL on two DT scenarios representing two network clusters in two different cities and compare its performance with the business-as-usual (BAU) policy and other competing optimization approaches using heuristic and Q-table algorithms. Benchmarking results show that CaRL performs the best and improves the average cluster-aggregated downlink throughput over the BAU policy by 24% and 18% in these two scenarios, respectively.

comment: 9 pages, 8 figures

♻ ☆ Energy-efficient UAV movement and user-UAV association in multi-UAV networks

Subhadip Ghosh, Priyadarshi Mukherjee, Sasthi C. Ghosh

These days, unmanned aerial vehicle (UAV)-based millimeter wave (mmWave) communication systems have drawn a lot of attention due to the increasing demand for faster data rates. Given the susceptibility of mmWave signals to obstacles and high propagation loss of mmWaves, ensuring line-of-sight (LoS) connectivity is critical for maintaining robust and efficient communication. Furthermore, UAVs have limited power resource and limited capacity in terms of number of users it can serve. Most significantly different users have different delay requirements and they keep moving while interacting with the UAVs. In this paper, first, we have provided an efficient solution for the optimal movement of the UAVs, by taking into account the energy efficiency of the UAVs as well as the mobility and delay priority of the users. Next, we have proposed a greedy solution for the optimal user-UAV assignment. After that, the numerical results show how well the suggested solution performs in comparison to the current benchmarks in terms of delay suffered by the users, number of unserved users, and energy efficiency of the UAVs.

comment: Submitted for a possible publication

♻ ☆ General Distribution Steering: A Sub-Optimal Solution by Convex Optimization

Guangyu Wu, Anders Lindquist

General distribution steering is intrinsically an infinite-dimensional problem, when the continuous distributions to steer are arbitrary. We put forward a moment representation of the primal system for control in [42]. However, the system trajectory was a predetermined one without optimization towards a design criterion, which doesn't always ensure a most satisfactory solution. In this paper, we propose an optimization approach to the general distribution steering problem of the first-order discrete-time linear system, i.e., an optimal control law for the corresponding moment system. The domain of all feasible control inputs is non-convex and has a complex topology. We obtain a subset of it by minimizing a weighted sum of squared integral distances alongside the system trajectory. The feasible domain is then proved convex, and the optimal control problem can be treated as a convex optimization or by exhaustive search, based on the type of the cost function. Algorithms of steering for continuous and discrete distributions are then put forward respectively, by adopting a realization scheme of control inputs. We also provide an explicit advantage of our proposed algorithm by truncated power moments to the prevailing Gaussian Mixture Models. Experiments on different types of cost functions are given to validate the performance of our proposed algorithm. Since the moment system is a dimension-reduced counterpart of the primal system, we call this solution a sub-optimal one to the primal general distribution steering problem.

comment: 16 pages, 23 figures

♻ ☆ A Minimal Control Family of Dynamical Systems for Universal Approximation

Yifei Duan, Yongqiang Cai

The universal approximation property (UAP) holds a fundamental position in deep learning, as it provides a theoretical foundation for the expressive power of neural networks. It is widely recognized that a composition of linear and nonlinear functions, such as the rectified linear unit (ReLU) activation function, can approximate continuous functions on compact domains. In this paper, we extend this efficacy to a scenario containing dynamical systems with controls. We prove that the control family $\mathcal{F}_1$ containing all affine maps and the nonlinear ReLU map is sufficient for generating flow maps that can approximate orientation-preserving (OP) diffeomorphisms on any compact domain. Since $\mathcal{F}_1$ contains only one nonlinear function and the UAP does not hold if we remove the nonlinear function, we call $\mathcal{F}_1$ a minimal control family for the UAP. On this basis, several mild sufficient conditions, such as affine invariance, are established for the control family and discussed. Our results reveal an underlying connection between the approximation power of neural networks and control systems and could provide theoretical guidance for examining the approximation power of flow-based models.

comment: 12 pages

♻ ☆ An Iterative Bayesian Approach for System Identification based on Linear Gaussian Models

Alexandros E. Tzikas, Mykel J. Kochenderfer

We tackle the problem of system identification, where we select inputs, observe the corresponding outputs from the true system, and optimize the parameters of our model to best fit the data. We propose a flexible and computationally tractable methodology that is compatible with any system and parametric family of models. Our approach only requires input-output data from the system and first-order information from the model with respect to the parameters. Our algorithm consists of two modules. First, we formulate the problem of system identification from a Bayesian perspective and use a linear Gaussian model approximation to iteratively optimize the model's parameters. In each iteration, we propose to use the input-output data to tune the covariance of the linear Gaussian model. This statistically calibrates the approach. Secondly, we define a Gaussian-based uncertainty measure for the model parameters, which we can then minimize with respect to the next selected input. We test our method with linear and nonlinear dynamics.

comment: Submitted to the IEEE CDC

♻ ☆ Insights into the explainability of Lasso-based DeePC for nonlinear systems

Gianluca Giacomelli, Simone Formentin, Victor G. Lopez, Matthias A. Müller, Valentina Breschi

Data-enabled Predictive Control (DeePC) has recently gained the spotlight as an easy-to-use control technique that allows for constraint handling while relying on raw data only. Initially proposed for linear time-invariant systems, several DeePC extensions are now available to cope with nonlinear systems. Nonetheless, these solutions mainly focus on ensuring the controller's effectiveness, overlooking the explainability of the final result. As a step toward explaining the outcome of DeePC for the control of nonlinear systems, in this paper, we focus on analyzing the earliest and simplest DeePC approach proposed to cope with nonlinearities in the controlled system, using a Lasso regularization. Our theoretical analysis highlights that the decisions undertaken by DeePC with Lasso regularization are unexplainable, as control actions are determined by data incoherent with the system's local behavior. This result is true even when the available input/output samples are grouped according to the different operating conditions explored during data collection. Our numerical study confirms these findings, highlighting the benefits of data grouping in terms of performance while showing that explainability remains a challenge in control design via DeePC.

♻ ☆ Set-point control and local stability for flat nonlinear systems using model-following control

Julian Willkomm, Kai Wulff, Johann Reger

We consider the set-point control problem for nonlinear systems with flat output that are subject to perturbations. The nonlinear dynamics as well as the perturbations are locally Lipschitz. We apply the model-following control (MFC) approach which consists of a model control loop (MCL) for a feedforward generation and a process control loop (PCL) that compensates the perturbations using high-gain feedback. We analyse the resulting closed-loop system and discuss its relation to a standard flatness-based high-gain approach. In particular we analyse the estimated region of attraction provided by a quadratic Lyapunov function. A case study illustrates the approach and quantifies the region of attraction obtained for each control approach. Using the initial condition of the model control loop as tuning parameter for the MFC design, provides that a significantly larger region of attraction can be guaranteed compared to a conventional single-loop high-gain design.

♻ ☆ Learning Algorithms for Verification of Markov Decision Processes

Tomáš Brázdil, Krishnendu Chatterjee, Martin Chmelik, Vojtěch Forejt, Jan Křetínský, Marta Kwiatkowska, Tobias Meggendorfer, David Parker, Mateusz Ujma

We present a general framework for applying learning algorithms and heuristical guidance to the verification of Markov decision processes (MDPs). The primary goal of our techniques is to improve performance by avoiding an exhaustive exploration of the state space, instead focussing on particularly relevant areas of the system, guided by heuristics. Our work builds on the previous results of Br{\'{a}}zdil et al., significantly extending it as well as refining several details and fixing errors. The presented framework focuses on probabilistic reachability, which is a core problem in verification, and is instantiated in two distinct scenarios. The first assumes that full knowledge of the MDP is available, in particular precise transition probabilities. It performs a heuristic-driven partial exploration of the model, yielding precise lower and upper bounds on the required probability. The second tackles the case where we may only sample the MDP without knowing the exact transition dynamics. Here, we obtain probabilistic guarantees, again in terms of both the lower and upper bounds, which provides efficient stopping criteria for the approximation. In particular, the latter is an extension of statistical model-checking (SMC) for unbounded properties in MDPs. In contrast to other related approaches, we do not restrict our attention to time-bounded (finite-horizon) or discounted properties, nor assume any particular structural properties of the MDP.

comment: 82 pages. This is the TheoretiCS journal version

♻ ☆ Optimal robust exact first-order differentiators with Lipschitz continuous output

Rodrigo Aldana-Lopez, Richard Seeber, Hernan Haimovich, David Gomez-Gutierrez

The signal differentiation problem involves the development of algorithms that allow to recover a signal's derivatives from noisy measurements. This paper develops a first-order differentiator with the following combination of properties: robustness to measurement noise, exactness in the absence of noise, optimal worst-case differentiation error, and Lipschitz continuous output where the output's Lipschitz constant is a tunable parameter. This combination of advantageous properties is not shared by any existing differentiator. Both continuous-time and sample-based versions of the differentiator are developed and theoretical guarantees are established for both. The continuous-time version of the differentiator consists in a regularized and sliding-mode-filtered linear adaptive differentiator. The sample-based, implementable version is then obtained through appropriate discretization. An illustrative example is provided to highlight the features of the developed differentiator.

♻ ☆ Emergent Cooperation for Energy-efficient Connectivity via Wireless Power Transfer

Winston Hurst, Anurag Pallaprolu, Yasamin Mostofi

This paper addresses the challenge of incentivizing energy-constrained, non-cooperative user equipment (UE) to serve as cooperative relays. We consider a source UE with a non-line-of-sight channel to an access point (AP), where direct communication may be infeasible or may necessitate a substantial transmit power. Other UEs in the vicinity are viewed as relay candidates, and our aim is to enable energy-efficient connectivity for the source, while accounting for the self-interested behavior and private channel state information of these candidates, by allowing the source to ``pay" the candidates via wireless power transfer (WPT). We propose a cooperation-inducing protocol, inspired by Myerson auction theory, which ensures that candidates truthfully report power requirements while minimizing the expected power used by the source. Through rigorous analysis, we establish the regularity of valuations for lognormal fading channels, which allows for the efficient determination of the optimal source transmit power. Extensive simulation experiments, employing real-world communication and WPT parameters, validate our theoretical framework. Our results demonstrate over 71% reduction in outage probability with as few as 4 relay candidates, compared to the non-cooperative scenario, and as much as 70% source power savings compared to a baseline approach, highlighting the efficacy of our proposed methodology.

♻ ☆ Uncrewed Vehicles in 6G Networks: A Unifying Treatment of Problems, Formulations, and Tools

Winston Hurst, Spilios Evmorfos, Athina Petropulu, Yasamin Mostofi

Uncrewed Vehicles (UVs) functioning as autonomous agents are anticipated to play a crucial role in the 6th Generation of wireless networks. Their seamless integration, cost-effectiveness, and the additional controllability through motion planning make them an attractive deployment option for a wide range of applications, both as assets in the network (e.g., mobile base stations) and as consumers of network services (e.g., autonomous delivery systems). However, despite their potential, the convergence of UVs and wireless systems brings forth numerous challenges that require attention from both academia and industry. This paper then aims to offer a comprehensive overview encompassing the transformative possibilities as well as the significant challenges associated with UV-assisted next-generation wireless communications. Considering the diverse landscape of possible application scenarios, problem formulations, and mathematical tools related to UV-assisted wireless systems, the underlying core theme of this paper is the unification of the problem space, providing a structured framework to understand the use cases, problem formulations, and necessary mathematical tools. Overall, the paper sets forth a clear understanding of how uncrewed vehicles can be integrated in the 6G ecosystem, paving the way towards harnessing the full potential at this intersection.

☆ Accelerated Approximate Optimization of Multi-Commodity Flows on Directed Graphs

Li Chen, Andrei Graur, Aaron Sidford

We provide $m^{1+o(1)}k\epsilon^{-1}$-time algorithms for computing multiplicative $(1 - \epsilon)$-approximate solutions to multi-commodity flow problems with $k$-commodities on $m$-edge directed graphs, including concurrent multi-commodity flow and maximum multi-commodity flow. To obtain our results, we provide new optimization tools of potential independent interest. First, we provide an improved optimization method for solving $\ell_{q, p}$-regression problems to high accuracy. This method makes $\tilde{O}_{q, p}(k)$ queries to a high accuracy convex minimization oracle for an individual block, where $\tilde{O}_{q, p}(\cdot)$ hides factors depending only on $q$, $p$, or $\mathrm{poly}(\log m)$, improving upon the $\tilde{O}_{q, p}(k^2)$ bound of [Chen-Ye, ICALP 2024]. As a result, we obtain the first almost-linear time algorithm that solves $\ell_{q, p}$ flows on directed graphs to high accuracy. Second, we present optimization tools to reduce approximately solving composite $\ell_{1, \infty}$-regression problems to solving $m^{o(1)}\epsilon^{-1}$ instances of composite $\ell_{q, p}$-regression problem. The method builds upon recent advances in solving box-simplex games [Jambulapati-Tian, NeurIPS 2023] and the area convex regularizer introduced in [Sherman, STOC 2017] to obtain faster rates for constrained versions of the problem. Carefully combining these techniques yields our directed multi-commodity flow algorithm.

☆ Faster Rates for No-Regret Learning in General Games via Cautious Optimism STOC 2025

Ashkan Soleymani, Georgios Piliouras, Gabriele Farina

We establish the first uncoupled learning algorithm that attains $O(n \log^2 d \log T)$ per-player regret in multi-player general-sum games, where $n$ is the number of players, $d$ is the number of actions available to each player, and $T$ is the number of repetitions of the game. Our results exponentially improve the dependence on $d$ compared to the $O(n\, d \log T)$ regret attainable by Log-Regularized Lifted Optimistic FTRL [Far+22c], and also reduce the dependence on the number of iterations $T$ from $\log^4 T$ to $\log T$ compared to Optimistic Hedge, the previously well-studied algorithm with $O(n \log d \log^4 T)$ regret [DFG21]. Our algorithm is obtained by combining the classic Optimistic Multiplicative Weights Update (OMWU) with an adaptive, non-monotonic learning rate that paces the learning process of the players, making them more cautious when their regret becomes too negative.

comment: Appeared at STOC 2025

☆ On Speedups for Convex Optimization via Quantum Dynamics

Shouvanik Chakrabarti, Dylan Herman, Jacob Watkins, Enrico Fontana, Brandon Augustino, Junhyung Lyle Kim, Marco Pistoia

We explore the potential for quantum speedups in convex optimization using discrete simulations of the Quantum Hamiltonian Descent (QHD) framework, as proposed by Leng et al., and establish the first rigorous query complexity bounds. We develop enhanced analyses for quantum simulation of Schr\"odinger operators with black-box potential via the pseudo-spectral method, providing explicit resource estimates independent of wavefunction assumptions. These bounds are applied to assess the complexity of optimization through QHD. Our findings pertain to unconstrained convex optimization in $d$ dimensions. In continuous time, we demonstrate that QHD, with suitable parameters, can achieve arbitrarily fast convergence rates. The optimization speed limit arises solely from the discretization of the dynamics, mirroring a property of the classical dynamics underlying QHD. Considering this cost, we show that a $G$-Lipschitz convex function can be optimized to an error of $\epsilon$ with $\widetilde{\mathcal{O}}(d^{1.5}G^2 R^2/\epsilon^2)$ queries. Moreover, under reasonable assumptions on the complexity of Hamiltonian simulation, $\widetilde{\Omega}(d/\epsilon^2)$ queries are necessary. Thus, QHD does not offer a speedup over classical zeroth order methods with exact oracles. However, we demonstrate that the QHD algorithm tolerates $\widetilde{\mathcal{O}}(\epsilon^3/d^{1.5}G^2 R^2)$ noise in function evaluation. We show that QHD offers a super-quadratic query advantage over all known classical algorithms tolerating this level of evaluation noise in the high-dimension regime. Additionally, we design a quantum algorithm for stochastic convex optimization that provides a super-quadratic speedup over all known classical algorithms in the high-dimension regime. To our knowledge, these results represent the first rigorous quantum speedups for convex optimization achieved through a dynamical algorithm.

☆ Linear Reweighted Regularization Algorithms for Graph Matching Problem

Rongxuan Li

The graph matching problem is a significant special case of the Quadratic Assignment Problem, with extensive applications in pattern recognition, computer vision, protein alignments and related fields. As the problem is NP-hard, relaxation and regularization techniques are frequently employed to improve tractability. However, most existing regularization terms are nonconvex, posing optimization challenges. In this paper, we propose a linear reweighted regularizer framework for solving the relaxed graph matching problem, preserving the convexity of the formulation. By solving a sequence of relaxed problems with the linear reweighted regularization term, one can obtain a sparse solution that, under certain conditions, theoretically aligns with the original graph matching problem's solution. Furthermore, we present a practical version of the algorithm by incorporating the projected gradient method. The proposed framework is applied to synthetic instances, demonstrating promising numerical results.

☆ Value of Information-based Deceptive Path Planning Under Adversarial Interventions

Wesley A. Suttle, Jesse Milzman, Mustafa O. Karabag, Brian M. Sadler, Ufuk Topcu

Existing methods for deceptive path planning (DPP) address the problem of designing paths that conceal their true goal from a passive, external observer. Such methods do not apply to problems where the observer has the ability to perform adversarial interventions to impede the path planning agent. In this paper, we propose a novel Markov decision process (MDP)-based model for the DPP problem under adversarial interventions and develop new value of information (VoI) objectives to guide the design of DPP policies. Using the VoI objectives we propose, path planning agents deceive the adversarial observer into choosing suboptimal interventions by selecting trajectories that are of low informational value to the observer. Leveraging connections to the linear programming theory for MDPs, we derive computationally efficient solution methods for synthesizing policies for performing DPP under adversarial interventions. In our experiments, we illustrate the effectiveness of the proposed solution method in achieving deceptiveness under adversarial interventions and demonstrate the superior performance of our approach to both existing DPP methods and conservative path planning approaches on illustrative gridworld problems.

comment: 10 pages, 4 figures

☆ A system level approach to generalised feedback Nash equilibrium seeking in partially-observed games

Otacilio B. L. Neto, Michela Mulas, Francesco Corona

This work proposes an algorithm for seeking generalised feedback Nash equilibria (GFNE) in noncooperative dynamic games. The focus is on cyber-physical systems with dynamics which are linear, stochastic, potentially unstable, and partially observed. We employ System Level Synthesis (SLS) to reformulate the problem as the search for an equilibrium profile of closed-loop responses to noise, which can then be used to reconstruct a stabilising output-feedback policy. Under this setup, we leverage monotone operator theory to design a GFNE-seeking algorithm capable to enforce closed-loop stability, operational constraints, and communication constraints onto the control policies. This algorithm is amenable to numerical implementation and we provide conditions for its convergence. We demonstrate our approach in a simulated experiment on the noncooperative stabilisation of a decentralised power-grid.

☆ Convexity of chance constraints for elliptical and skewed distributions with copula structures dependent on decision variables

Heng Zhang, Abdel Lisser

Chance constraints describe a set of given random inequalities depending on the decision vector satisfied with a large enough probability. They are widely used in decision making under uncertain data in many engineering problems. This paper aims to derive the convexity of chance constraints with row dependent elliptical and skewed random variables via a copula depending on decision vectors. We obtain best thresholds of the $r$-concavity for any real number $r$ and improve probability thresholds of the eventual convexity. We prove the eventual convexity with elliptical distributions and a Gumbel-Hougaard copula despite the copula's singularity near the origin. We determine the $\alpha$-decreasing densities of generalized hyperbolic distributions by estimating the modified Bessel functions. By applying the $\alpha$-decreasing property and a radial decomposition, we achieve the eventual convexity for three types of skewed distributions. Finally, we provide an example to illustrate the eventual convexity of a feasible set containing the origin.

comment: 64 pages, 6 figures

☆ Robust Feedback Optimization with Model Uncertainty: A Regularization Approach

Winnie Chan, Zhiyu He, Keith Moffat, Saverio Bolognani, Michael Muehlebach, Florian Dörfler

Feedback optimization optimizes the steady state of a dynamical system by implementing optimization iterations in closed loop with the plant. It relies on online measurements and limited model information, namely, the input-output sensitivity. In practice, various issues including inaccurate modeling, lack of observation, or changing conditions can lead to sensitivity mismatches, causing closed-loop sub-optimality or even instability. To handle such uncertainties, we pursue robust feedback optimization, where we optimize the closed-loop performance against all possible sensitivities lying in specific uncertainty sets. We provide tractable reformulations for the corresponding min-max problems via regularizations and characterize the online closed-loop performance through the tracking error in case of time-varying optimal solutions. Simulations on a distribution grid illustrate the effectiveness of our robust feedback optimization controller in addressing sensitivity mismatches in a non-stationary environment.

☆ The Influence of an Adjoint Mismatch on the Primal-Dual Douglas-Rachford Method

Emanuele Naldi, Felix Schneppe

The primal-dual Douglas-Rachford method is a well-known algorithm to solve optimization problems written as convex-concave saddle-point problems. Each iteration involves solving a linear system involving a linear operator and its adjoint. However, in practical applications it is often computationally favorable to replace the adjoint operator by a computationally more efficient approximation. This leads to an adjoint mismatch. In this paper, we analyze the convergence of the primal-dual Douglas-Rachford method under the presence of an adjoint mismatch. We provide mild conditions that guarantee the existence of a fixed point and find an upper bound on the error of the primal solution. Furthermore, we establish step sizes in the strongly convex setting that guarantee linear convergence under mild conditions. Additionally, we provide an alternative method that can also be derived from the Douglas-Rachford method and is also guaranteed to converge in this setting. Moreover, we illustrate our results both for an academic and a real-world inspired example.

comment: 37 pages, 10 figures

☆ Forward-backward splitting in bilaterally bounded Alexandrov spaces

Heikki von Koch, Tuomo Valkonen

With the goal of solving optimisation problems on non-Riemannian manifolds, such as geometrical surfaces with sharp edges, we develop and prove the convergence of a forward-backward method in Alexandrov spaces with curvature bounded both from above and from below. This bilateral boundedness is crucial for the availability of both the gradient and proximal steps, instead of just one or the other. We numerically demonstrate the behaviour of the proposed method on simple geometrical surfaces in $\mathbb{R}^3$.

☆ Application of Battery Storage to Switching Predictive Control of Power Distribution Systems Including Road Heating

Chiaki Kojima, Yuya Muto, Hikaru Akutsu, Rinnosuke Shima, Yoshihiko Susuki

A road heating system is an electrical device which promotes snow melting by burying a heating cable as a thermal source underground. When integrating road heating into the power distribution system, we need to optimize the flow of electric power by appropriately integrating distributed power sources and conventional power distribution equipment. In this paper, we extend the power distribution system considered in the authors' previous study to the case where battery storage is installed. As a main result, we propose a predictive switching control that achieves the reduction of distribution loss, attenuation of voltage fluctuation, and efficient snow melting, simultaneously. We verify the effectiveness of the application of battery storage through numerical simulation.

comment: 13 pages, 14 figures

☆ Distributed AC Optimal Power Flow: A Scalable Solution for Large-Scale Problems

Xinliang Dai, Yuning Jiang, Yi Guo, Colin N. Jones, Moritz Diehl, Veit Hagenmeyer

This paper introduces a novel distributed optimization framework for large-scale AC Optimal Power Flow (OPF) problems, offering both theoretical convergence guarantees and rapid convergence in practice. By integrating smoothing techniques and the Schur complement, the proposed approach addresses the scalability challenges and reduces communication overhead in distributed AC OPF. Additionally, optimal network decomposition enables efficient parallel processing under the single program multiple data (SPMD) paradigm. Extensive simulations on large-scale benchmarks across various operating scenarios indicate that the proposed framework outperforms the state-of-the-art centralized solver IPOPT on modest hardware. This paves the way for more scalable and efficient distributed optimization in future power system applications.

☆ Riemannian Multiplicative Update for Sparse Simplex constraint using oblique rotation manifold

Flavia Esposito, Andersen Ang

We propose a new manifold optimization method to solve low-rank problems with sparse simplex constraints (variables are simultaneous nonnegativity, sparsity, and sum-to-1) that are beneficial in applications. The proposed approach exploits oblique rotation manifolds, rewrite the problem, and introduce a new Riemannian optimization method. Experiments on synthetic datasets compared to the standard Euclidean method show the effectiveness of the proposed method.

comment: 8 pages, 1 figure

☆ Bi-Level Route Optimization and Path Planning with Hazard Exploration

Jimin Choi, Grant Stagg, Cameron K. Peterson, Max Z. Li

Effective risk monitoring in dynamic environments such as disaster zones requires an adaptive exploration strategy to detect hidden threats. We propose a bi-level unmanned aerial vehicle (UAV) monitoring strategy that efficiently integrates high-level route optimization with low-level path planning for known and unknown hazards. At the high level, we formulate the route optimization as a vehicle routing problem (VRP) to determine the optimal sequence for visiting known hazard locations. To strategically incorporate exploration efficiency, we introduce an edge-based centroidal Voronoi tessellation (CVT), which refines baseline routes using pseudo-nodes and allocates path budgets based on the UAV's battery capacity using a line segment Voronoi diagram. At the low level, path planning maximizes information gain within the allocated path budget by generating kinematically feasible B-spline trajectories. Bayesian inference is applied to dynamically update hazard probabilities, enabling the UAVs to prioritize unexplored regions. Simulation results demonstrate that edge-based CVT improves spatial coverage and route uniformity compared to the node-based method. Additionally, our optimized path planning consistently outperforms baselines in hazard discovery rates across a diverse set of scenarios.

☆ An ANN-Enhanced Approach for Flatness-Based Constrained Control of Nonlinear Systems

Huu-Thinh Do, Ionela Prodan, Florin Stoican

Neural networks have proven practical for a synergistic combination of advanced control techniques. This work analyzes the implementation of rectified linear unit neural networks to achieve constrained control in differentially flat systems. Specifically, the class of flat systems enjoys the benefit of feedback linearizability, i.e., the systems can be linearized by means of a proper variable transformation. However, the price for linearizing the dynamics is that the constraint descriptions are distorted geometrically. Our results show that, by using neural networks, these constraints can be represented as a union of polytopes, enabling the use of mixed-integer programming tools to guarantee constraint satisfaction. We further analyze the integration of the characterization into efficient settings such as control Lyapunov function-based and model predictive control (MPC). Interestingly, this description also allows us to explicitly compute the solution of the MPC problem for the nonlinear system. Several examples are provided to illustrate the effectiveness of our framework.

☆ Consensus on Open Multi-Agent Systems Over Graphs Sampled from Graphons

Renato Vizuete, Julien M. Hendrickx

We show how graphons can be used to model and analyze open multi-agent systems, which are multi-agent systems subject to arrivals and departures, in the specific case of linear consensus. First, we analyze the case of replacements, where under the assumption of a deterministic interval between two replacements, we derive an upper bound for the disagreement in expectation. Then, we study the case of arrivals and departures, where we define a process for the evolution of the number of agents that guarantees a minimum and a maximum number of agents. Next, we derive an upper bound for the disagreement in expectation, and we establish a link with the spectrum of the expected graph used to generate the graph topologies. Finally, for stochastic block model (SBM) graphons, we prove that the computation of the spectrum of the expected graph can be performed based on a matrix whose dimension depends only on the graphon and it is independent of the number of agents.

comment: 8 pages, 1 figure

☆ Deviation Between Team-Optimal Solution and Nash Equilibrium in Flow Assignment Problems

Gehui Xu, Ting Bai, Andreas A. Malikopoulos, Thomas Parisini

We investigate the relationship between the team-optimal solution and the Nash equilibrium (NE) to assess the impact of strategy deviation on team performance. As a working use case, we focus on a class of flow assignment problems in which each source node acts as a cooperating decision maker (DM) within a team that minimizes the team cost based on the team-optimal strategy. In practice, some selfish DMs may prioritize their own marginal cost and deviate from NE strategies, thus potentially degrading the overall performance. To quantify this deviation, we explore the deviation bound between the team-optimal solution and the NE in two specific scenarios: (i) when the team-optimal solution is unique and (ii) when multiple solutions do exist. This helps DMs analyze the factors influencing the deviation and adopting the NE strategy within a tolerable range. Furthermore, in the special case of a potential game model, we establish the consistency between the team-optimal solution and the NE. Once the consistency condition is satisfied, the strategy deviation does not alter the total cost, and DMs do not face a strategic trade-off. Finally, we validate our theoretical analysis through some simulation studies.

☆ Federated Structured Sparse PCA for Anomaly Detection in IoT Networks

Chenyi Huang, Xinrong Li, Xianchao Xiu

Although federated learning has gained prominence as a privacy-preserving framework tailored for distributed Internet of Things (IoT) environments, current federated principal component analysis (PCA) methods lack integration of sparsity, a critical feature for robust anomaly detection. To address this limitation, we propose a novel federated structured sparse PCA (FedSSP) approach for anomaly detection in IoT networks. The proposed model uniquely integrates double sparsity regularization: (1) row-wise sparsity governed by $\ell_{2,p}$-norm with $p\in[0,1)$ to eliminate redundant feature dimensions, and (2) element-wise sparsity via $\ell_{q}$-norm with $q\in[0,1)$ to suppress noise-sensitive components. To efficiently solve this non-convex optimization problem in a distributed setting, we devise a proximal alternating minimization (PAM) algorithm with rigorous theoretical proofs establishing its convergence guarantees. Experiments on real datasets validate that incorporating structured sparsity enhances both model interpretability and detection accuracy.

☆ Distributionally Robust Model Order Reduction for Linear Systems

Le Liu, Yu Kawano, Yangming Dou, Ming Cao

In this paper, we investigate distributionally robust model order reduction for linear, discrete-time, time-invariant systems. The external input is assumed to follow an uncertain distribution within a Wasserstein ambiguity set. We begin by considering the case where the distribution is certain and formulate an optimization problem to obtain the reduced model. When the distribution is uncertain, the interaction between the reduced-order model and the distribution is modeled by a Stackelberg game. To ensure solvability, we first introduce the Gelbrich distance and demonstrate that the Stackelberg game within a Wasserstein ambiguity set is equivalent to that within a Gelbrich ambiguity set. Then, we propose a nested optimization problem to solve the Stackelberg game. Furthermore, the nested optimization problem is relaxed into a nested convex optimization problem, ensuring computational feasibility. Finally, a simulation is presented to illustrate the effectiveness of the proposed method.

☆ Certified Approximate Reachability (CARe): Formal Error Bounds on Deep Learning of Reachable Sets

Prashant Solanki, Nikolaus Vertovec, Yannik Schnitzer, Jasper Van Beers, Coen de Visser, Alessandro Abate

Recent approaches to leveraging deep learning for computing reachable sets of continuous-time dynamical systems have gained popularity over traditional level-set methods, as they overcome the curse of dimensionality. However, as with level-set methods, considerable care needs to be taken in limiting approximation errors, particularly since no guarantees are provided during training on the accuracy of the learned reachable set. To address this limitation, we introduce an epsilon-approximate Hamilton-Jacobi Partial Differential Equation (HJ-PDE), which establishes a relationship between training loss and accuracy of the true reachable set. To formally certify this approximation, we leverage Satisfiability Modulo Theories (SMT) solvers to bound the residual error of the HJ-based loss function across the domain of interest. Leveraging Counter Example Guided Inductive Synthesis (CEGIS), we close the loop around learning and verification, by fine-tuning the neural network on counterexamples found by the SMT solver, thus improving the accuracy of the learned reachable set. To the best of our knowledge, Certified Approximate Reachability (CARe) is the first approach to provide soundness guarantees on learned reachable sets of continuous dynamical systems.

☆ Sampled-data and event-triggered control of globally Lipschitz infinite-dimensional systems

Rami Katz, Andrii Mironchenko

We show that if a linear infinite-dimensional system is exponentially stabilizable by compact feedback, it is also stabilizable by means of a sampled-data feedback that is fed through a globally Lipschitz nonlinearity, provided that the sector bound for the nonlinearity and the sampling time is small enough. Next we develop a switching-based event-triggered control scheme stabilizing the system with a reduced number of switching events. We demonstrate our results on an example of finite-dimensional stabilization of a Sturm-Liouville parabolic system.

☆ An extrapolated and provably convergent algorithm for nonlinear matrix decomposition with the ReLU function

Nicolas Gillis, Margherita Porcelli, Giovanni Seraghiti

Nonlinear matrix decomposition (NMD) with the ReLU function, denoted ReLU-NMD, is the following problem: given a sparse, nonnegative matrix $X$ and a factorization rank $r$, identify a rank-$r$ matrix $\Theta$ such that $X\approx \max(0,\Theta)$. This decomposition finds application in data compression, matrix completion with entries missing not at random, and manifold learning. The standard ReLU-NMD model minimizes the least squares error, that is, $\|X - \max(0,\Theta)\|_F^2$. The corresponding optimization problem is nondifferentiable and highly nonconvex. This motivated Saul to propose an alternative model, Latent-ReLU-NMD, where a latent variable $Z$ is introduced and satisfies $\max(0,Z)=X$ while minimizing $\|Z - \Theta\|_F^2$ (``A nonlinear matrix decomposition for mining the zeros of sparse data'', SIAM J. Math. Data Sci., 2022). Our first contribution is to show that the two formulations may yield different low-rank solutions $\Theta$; in particular, we show that Latent-ReLU-NMD can be ill-posed when ReLU-NMD is not, meaning that there are instances in which the infimum of Latent-ReLU-NMD is not attained while that of ReLU-NMD is. We also consider another alternative model, called 3B-ReLU-NMD, which parameterizes $\Theta=WH$, where $W$ has $r$ columns and $H$ has $r$ rows, allowing one to get rid of the rank constraint in Latent-ReLU-NMD. Our second contribution is to prove the convergence of a block coordinate descent (BCD) applied to 3B-ReLU-NMD and referred to as BCD-NMD. Our third contribution is a novel extrapolated variant of BCD-NMD, dubbed eBCD-NMD, which we prove is also convergent under mild assumptions. We illustrate the significant acceleration effect of eBCD-NMD compared to BCD-NMD, and also show that eBCD-NMD performs well against the state of the art on synthetic and real-world data sets.

comment: 27 pages. Codes and data available from https://github.com/giovanniseraghiti/ReLU-NMD

☆ Shannon-and Von Neumann-entropy regularizations of linear and semidefinite programs

Saroj Prasad Chhatoi, Jean B Lasserre

We consider the LP in standard form min {c T x : Ax = b; x $\ge$ 0} and inspired by $\epsilon$-regularization in Optimal Transport, we introduce its $\epsilon$-regularization ''min {c T x + $\epsilon$ f (x) : Ax = b; x $\ge$ 0}'' via the (convex) Boltzmann-Shannon entropy f (x) := i x i ln x i . We also provide a similar regularization for the semidefinite program ''min {Tr(C $\bullet$ X) : A(X) = b; X 0}'' but with now the so-called Von Neumann entropy, as in Quantum Optimal Transport. Importantly, both are not barriers of the LP and SDP cones respectively. We show that this problem admits an equivalent unconstrained convex problem max $\lambda$$\in$R m G$\epsilon$($\lambda$) for an explicit concave differentiable function G$\epsilon$ in dual variables $\lambda$ $\in$ R m . As $\epsilon$ goes to zero, its optimal value converges to the optimal value of the initial LP. While it resembles the log-barrier formulation of interior point algorithm for the initial LP, it has a distinguishing advantage. Namely for fixed $\lambda$, G$\epsilon$($\lambda$) is obtained as a minimization over the whole space x $\in$ R d (and not over x $\ge$ 0) to still obtain a nonnegative solution x($\lambda$) $\ge$ 0, whence an explicit form of G$\epsilon$ very useful for its unconstrained maximization over R m.

☆ Joint Replenishment Strategy for Multiple Satellite Constellations with Shared Launch Opportunities

Jaewoo Kim, Taehyun Sung, Woonam Hwang, Jaemyung Ahn

This paper proposes a novel replenishment strategy that can jointly support multiple satellite constellations. In this approach, multiple constellations share launch opportunities and parking orbits to address the operational satellite failures and ensure the desired service level of the constellations. We develop an inventory management model based on parametric replenishment policies, considering the launch vehicle's capacity and the shipping size of satellites. Based on this model, we introduce two decision-making scenarios and propose their corresponding solution frameworks. We conduct two case studies to provide valuable insights into the proposed strategy and demonstrate its applicability to supply chain management for maintaining multiple satellite constellations.

☆ A Decomposition Approach for the Gain Function in the Feedback Particle Filter

Ruoyu Wang, Huimin Miao, Xue Luo

The feedback particle filter (FPF) is an innovative, control-oriented and resampling-free adaptation of the traditional particle filter (PF). In the FPF, individual particles are regulated via a feedback gain, and the corresponding gain function serves as the solution to the Poisson's equation equipped with a probability-weighted Laplacian. Owing to the fact that closed-form expressions can only be computed under specific circumstances, approximate solutions are typically indispensable. This paper is centered around the development of a novel algorithm for approximating the gain function in the FPF. The fundamental concept lies in decomposing the Poisson's equation into two equations that can be precisely solved, provided that the observation function is a polynomial. A free parameter is astutely incorporated to guarantee exact solvability. The computational complexity of the proposed decomposition method shows a linear correlation with the number of particles and the polynomial degree of the observation function. We perform comprehensive numerical comparisons between our method, the PF, and the FPF using the constant-gain approximation and the kernel-based approach. Our decomposition method outperforms the PF and the FPF with constant-gain approximation in terms of accuracy. Additionally, it has the shortest CPU time among all the compared methods with comparable performance.

comment: 8 pages, 5 figures, submitted to IEEE CDC

☆ Remarks on the Polyak-Lojasiewicz inequality and the convergence of gradient systems

Arthur Castello B. de Oliveira, Leilei Cui, Eduardo D. Sontag

This work explores generalizations of the Polyak-Lojasiewicz inequality (PLI) and their implications for the convergence behavior of gradient flows in optimization problems. Motivated by the continuous-time linear quadratic regulator (CT-LQR) policy optimization problem -- where only a weaker version of the PLI is characterized in the literature -- this work shows that while weaker conditions are sufficient for global convergence to, and optimality of the set of critical points of the cost function, the "profile" of the gradient flow solution can change significantly depending on which "flavor" of inequality the cost satisfies. After a general theoretical analysis, we focus on fitting the CT-LQR policy optimization problem to the proposed framework, showing that, in fact, it can never satisfy a PLI in its strongest form. We follow up our analysis with a brief discussion on the difference between continuous- and discrete-time LQR policy optimization, and end the paper with some intuition on the extension of this framework to optimization problems with L1 regularization and solved through proximal gradient flows.

♻ ☆ Revisiting Lossless Convexification: Theoretical Guarantees for Discrete-time Optimal Control Problems

Dayou Luo, Kazuya Echigo, Behçet Açıkmeşe

Lossless Convexification (LCvx) is a modeling approach that transforms a class of nonconvex optimal control problems, where nonconvexity primarily arises from control constraints, into convex problems through convex relaxations. These convex problems can be solved using polynomial-time numerical methods after discretization, which converts the original infinite-dimensional problem into a finite-dimensional one. However, existing LCvx theory is limited to continuous-time optimal control problems, as the equivalence between the relaxed convex problem and the original nonconvex problem holds only in continuous time. This paper extends LCvx to discrete-time optimal control problems by classifying them into normal and long-horizon cases. For normal cases, after an arbitrarily small perturbation to the system dynamics (recursive equality constraints), applying the existing LCvx method to discrete-time problems results in optimal controls that meet the original nonconvex constraints at all but no more than $n_x - 1$ temporal grid points, where $n_x$ is the state dimension. For long-horizon cases, the existing LCvx method fails, but we resolve this issue by integrating it with a bisection search, leveraging the continuity of the value function from the relaxed convex problem to achieve similar results as in normal cases. This paper improves the theoretical foundation of LCvx, expanding its applicability to real-world discrete-time optimal control problems.

♻ ☆ Diffusion at Absolute Zero: Langevin Sampling using Successive Moreau Envelopes

Andreas Habring, Alexander Falk, Martin Zach, Thomas Pock

We propose a method for sampling from Gibbs distributions of the form $\pi(x)\propto\exp(-U(x))$ by considering a family $(\pi^{t})_t$ of approximations of the target density which is such that $\pi^{t}$ exhibits favorable properties for sampling when $t$ is large, and $\pi^{t} \to \pi$ as $t \to 0$. This sequence is obtained by replacing (parts of) the potential $U$ by its Moreau envelope. Through the sequential sampling from $\pi^{t}$ for decreasing values of $t$ by a Langevin algorithm with appropriate step size, the samples are guided from a simple starting density to the more complex target quickly. We prove the ergodicity of the method as well as its convergence to the target density without assuming convexity or differentiability of the potential $U$. In addition to the theoretical analysis, we show experimental results that support the superiority of the method in terms of convergence speed and mode-coverage of multi-modal densities to current algorithms. The experiments range from one-dimensional toy-problems to high-dimensional inverse imaging problems with learned potentials.

♻ ☆ Simple matrix expressions for the curvatures of Grassmannian

Zehua Lai, Lek-Heng Lim, Ke Ye

We show that modeling a Grassmannian as symmetric orthogonal matrices $\operatorname{Gr}(k,\mathbb{R}^n) \cong\{Q \in \mathbb{R}^{n \times n} : Q^{\scriptscriptstyle\mathsf{T}} Q = I, \; Q^{\scriptscriptstyle\mathsf{T}} = Q,\; \operatorname{tr}(Q)=2k - n\}$ yields exceedingly simple matrix formulas for various curvatures and curvature-related quantities, both intrinsic and extrinsic. These include Riemann, Ricci, Jacobi, sectional, scalar, mean, principal, and Gaussian curvatures; Schouten, Weyl, Cotton, Bach, Pleba\'nski, cocurvature, nonmetricity, and torsion tensors; first, second, and third fundamental forms; Gauss and Weingarten maps; and upper and lower delta invariants. We will derive explicit, simple expressions for the aforementioned quantities in terms of standard matrix operations that are stably computable with numerical linear algebra. Many of these aforementioned quantities have never before been presented for the Grassmannian.

comment: 26 pages

♻ ☆ Feasibility Evaluation of Quadratic Programs for Constrained Control

Panagiotis Rousseas, Dimitra Panagou

This paper presents a computationally-efficient method for evaluating the feasibility of Quadratic Programs (QPs) for online constrained control. Based on the duality principle, we first show that the feasibility of a QP can be determined by the solution of a properly-defined Linear Program (LP). Our analysis yields a LP that can be solved more efficiently compared to the original QP problem, and more importantly, is simpler in form and can be solved more efficiently compared to existing methods that assess feasibility via LPs. The computational efficiency of the proposed method compared to existing methods for feasibility evaluation is demonstrated in comparative case studies as well as a feasible-constraint selection problem, indicating its promise for online feasibility evaluation of optimization-based controllers.

comment: Submitted to CDC 2025

♻ ☆ A Framework for Approximating Perturbed Optimal Control Problems

Riley Link, Ethan Ebbighausen

We consider trajectory optimal control problems in which parameter uncertainty limits the applicability of control trajectories computed prior to travel. Hence, efficient trajectory adjustment is needed to ensure successful travel. However, it is often prohibitive or impossible to recalculate the optimal control in-transit due to strict time constraints or limited onboard computing resources. Thus, we propose a framework for quick and accurate trajectory approximations by using post-optimality sensitivity information. This allows the reduction of uncertain parameter space and an instantaneous approximation of the new optimal controller while using sensitivity data computed and stored pretransit.

♻ ☆ Distributed Fractional Bayesian Learning for Adaptive Optimization

Yaqun Yang, Jinlong Lei, Guanghui Wen, Yiguang Hong

This paper considers a distributed adaptive optimization problem, where all agents only have access to their local cost functions with a common unknown parameter, whereas they mean to collaboratively estimate the true parameter and find the optimal solution over a connected network. A general mathematical framework for such a problem has not been studied yet. We aim to provide valuable insights for addressing parameter uncertainty in distributed optimization problems and simultaneously find the optimal solution. Thus, we propose a novel Prediction while Optimization scheme, which utilizes distributed fractional Bayesian learning through weighted averaging on the log-beliefs to update the beliefs of unknown parameters, and distributed gradient descent for renewing the estimation of the optimal solution. Then under suitable assumptions, we prove that all agents' beliefs and decision variables converge almost surely to the true parameter and the optimal solution under the true parameter, respectively. We further establish a sublinear convergence rate for the belief sequence. Finally, numerical experiments are implemented to corroborate the theoretical analysis.

♻ ☆ Efficient Learning for Entropy-Regularized Markov Decision Processes via Multilevel Monte Carlo

Matthieu Meunier, Christoph Reisinger, Yufei Zhang

Designing efficient learning algorithms with complexity guarantees for Markov decision processes (MDPs) with large or continuous state and action spaces remains a fundamental challenge. We address this challenge for entropy-regularized MDPs with Polish state and action spaces, assuming access to a generative model of the environment. We propose a novel family of multilevel Monte Carlo (MLMC) algorithms that integrate fixed-point iteration with MLMC techniques and a generic stochastic approximation of the Bellman operator. We quantify the precise impact of the chosen approximate Bellman operator on the accuracy of the resulting MLMC estimator. Leveraging this error analysis, we show that using a biased plain MC estimate for the Bellman operator results in quasi-polynomial sample complexity, whereas an unbiased randomized multilevel approximation of the Bellman operator achieves polynomial sample complexity in expectation. Notably, these complexity bounds are independent of the dimensions or cardinalities of the state and action spaces, distinguishing our approach from existing algorithms whose complexities scale with the sizes of these spaces. We validate these theoretical performance guarantees through numerical experiments.

comment: 46 pages, 6 figures; fixed formatting of definitions and titles

♻ ☆ Stability of Schrödinger bridges and Sinkhorn semigroups for log-concave models

Pierre Del Moral

In this article we obtain several new results and developments in the study of entropic optimal transport problems (a.k.a. Schr\"odinger problems) with general reference distributions and log-concave target marginal measures. Our approach combines transportation cost inequalities with the theory of Riccati matrix difference equations arising in filtering and optimal control theory. This methodology is partly based on a novel entropic stability of Schr\"odinger bridges and closed form expressions of a class of discrete time algebraic Riccati equations. In the context of regularized entropic transport these techniques provide new sharp entropic map estimates. When applied to the stability of Sinkhorn semigroups, they also yield a series of novel contraction estimates in terms of the fixed point of Riccati equations. The strength of our approach is that it is applicable to a large class of models arising in machine learning and artificial intelligence algorithms. We illustrate the impact of our results in the context of regularized entropic transport, proximal samplers and diffusion generative models as well as diffusion flow matching models

♻ ☆ Implicit Bias and Fast Convergence Rates for Self-attention

Bhavya Vasudeva, Puneesh Deora, Christos Thrampoulidis

We study the fundamental optimization principles of self-attention, the defining mechanism of transformers, by analyzing the implicit bias of gradient-based optimizers in training a self-attention layer with a linear decoder in binary classification. Building on prior studies in linear logistic regression, recent findings demonstrate that the key-query matrix $W_t$ from gradient-descent (GD) converges in direction towards $W_{mm}$, which maximizes the margin between optimal and non-optimal tokens across sequences. However, this convergence is local, dependent on initial conditions, only holds asymptotically as the number of iterations increases, and leaves questions about the potential benefits of adaptive step-size rules unaddressed. To bridge this gap, we first establish scenarios for which convergence is provably \emph{global}. We then analyze two adaptive step-size strategies: normalized GD and Polyak step-size, demonstrating \emph{finite-time} convergence rates for $W_t$ to $W_{mm}$, and quantifying the sparsification rate of the attention map. These findings not only show that these strategies can accelerate parameter convergence over standard GD in a non-convex setting but also deepen the understanding of the implicit bias in self-attention, linking it more closely to the phenomena observed in linear logistic regression despite its intricate non-convex nature.

comment: Accepted in TMLR, 43 pages, 10 figures

♻ ☆ ADMM Algorithms for Residual Network Training: Convergence Analysis and Parallel Implementation

Jintao Xu, Yifei Li, Wenxun Xing

We propose both serial and parallel proximal (linearized) alternating direction method of multipliers (ADMM) algorithms for training residual neural networks. In contrast to backpropagation-based approaches, our methods inherently mitigate the exploding gradient issue and are well-suited for parallel and distributed training through regional updates. Theoretically, we prove that the proposed algorithms converge at an R-linear (sublinear) rate for both the iteration points and the objective function values. These results hold without imposing stringent constraints on network width, depth, or training data size. Furthermore, we theoretically analyze our parallel/distributed ADMM algorithms, highlighting their reduced time complexity and lower per-node memory consumption. To facilitate practical deployment, we develop a control protocol for parallel ADMM implementation using Python's multiprocessing and interprocess communication. Experimental results validate the proposed ADMM algorithms, demonstrating rapid and stable convergence, improved performance, and high computational efficiency. Finally, we highlight the improved scalability and efficiency achieved by our parallel ADMM training strategy.

♻ ☆ General Distribution Steering: A Sub-Optimal Solution by Convex Optimization

Guangyu Wu, Anders Lindquist

General distribution steering is intrinsically an infinite-dimensional problem, when the continuous distributions to steer are arbitrary. We put forward a moment representation of the primal system for control in [42]. However, the system trajectory was a predetermined one without optimization towards a design criterion, which doesn't always ensure a most satisfactory solution. In this paper, we propose an optimization approach to the general distribution steering problem of the first-order discrete-time linear system, i.e., an optimal control law for the corresponding moment system. The domain of all feasible control inputs is non-convex and has a complex topology. We obtain a subset of it by minimizing a weighted sum of squared integral distances alongside the system trajectory. The feasible domain is then proved convex, and the optimal control problem can be treated as a convex optimization or by exhaustive search, based on the type of the cost function. Algorithms of steering for continuous and discrete distributions are then put forward respectively, by adopting a realization scheme of control inputs. We also provide an explicit advantage of our proposed algorithm by truncated power moments to the prevailing Gaussian Mixture Models. Experiments on different types of cost functions are given to validate the performance of our proposed algorithm. Since the moment system is a dimension-reduced counterpart of the primal system, we call this solution a sub-optimal one to the primal general distribution steering problem.

comment: 16 pages, 23 figures

♻ ☆ Distributed Optimization Method Based On Optimal Control

Ziyuan Guo, Yue Sun, Yeming Xu, Liping Zhang, Huanshui Zhang

In this paper, a novel distributed optimization framework has been proposed. The key idea is to convert optimization problems into optimal control problems where the objective of each agent is to design the current control input minimizing the original objective function of itself and updated size for the future time instant. Compared with the existing distributed optimization problem for optimizing a sum of convex objective functions corresponding to multiple agents, we present a distributed optimization algorithm for multi-agents system based on the results from the maximum principle. Moreover, the convergence and superlinear convergence rate are also analyzed stringently.

♻ ☆ ADMM for Structured Fractional Minimization

Ganzhao Yuan

This paper considers a class of structured fractional minimization problems. The numerator consists of a differentiable function, a simple nonconvex nonsmooth function, a concave nonsmooth function, and a convex nonsmooth function composed with a linear operator. The denominator is a continuous function that is either weakly convex or has a weakly convex square root. These problems are prevalent in various important applications in machine learning and data science. Existing methods, primarily based on subgradient methods and smoothing proximal gradient methods, often suffer from slow convergence and numerical stability issues. In this paper, we introduce {\sf FADMM}, the first Alternating Direction Method of Multipliers tailored for this class of problems. {\sf FADMM} decouples the original problem into linearized proximal subproblems, featuring two variants: one using Dinkelbach's parametric method ({\sf FADMM-D}) and the other using the quadratic transform method ({\sf FADMM-Q}). By introducing a novel Lyapunov function, we establish that {\sf FADMM} converges to $\epsilon$-approximate critical points of the problem within an oracle complexity of $\mathcal{O}(1/\epsilon^{3})$. Extensive experiments on synthetic and real-world datasets, including sparse Fisher discriminant analysis, robust Sharpe ratio minimization, and robust sparse recovery, demonstrate the effectiveness of our approach. Keywords: Fractional Minimization, Nonconvex Optimization, Proximal Linearized ADMM, Nonsmooth Optimization, Convergence Analysis

♻ ☆ ADMM for Nonconvex Optimization under Minimal Continuity Assumption

Ganzhao Yuan

This paper introduces a novel approach to solving multi-block nonconvex composite optimization problems through a proximal linearized Alternating Direction Method of Multipliers (ADMM). This method incorporates an Increasing Penalization and Decreasing Smoothing (IPDS) strategy. Distinguishing itself from existing ADMM-style algorithms, our approach (denoted IPDS-ADMM) imposes a less stringent condition, specifically requiring continuity in just one block of the objective function. IPDS-ADMM requires that the penalty increases and the smoothing parameter decreases, both at a controlled pace. When the associated linear operator is bijective, IPDS-ADMM uses an over-relaxation stepsize for faster convergence; however, when the linear operator is surjective, IPDS-ADMM uses an under-relaxation stepsize for global convergence. We devise a novel potential function to facilitate our convergence analysis and prove an oracle complexity $\mathcal{O}(\epsilon^{-3})$ to achieve an $\epsilon$-approximate critical point. To the best of our knowledge, this is the first complexity result for using ADMM to solve this class of nonsmooth nonconvex problems. Finally, some experiments on the sparse PCA problem are conducted to demonstrate the effectiveness of our approach.

♻ ☆ Superlinear Optimization Algorithms

Hongxia Wang, Yeming Xu, Ziyuan Guo, Huanshui Zhang

This paper proposes several novel optimization algorithms for minimizing a nonlinear objective function. The algorithms are enlightened by the optimal state trajectory of an optimal control problem closely related to the minimized objective function. They are superlinear convergent when appropriate parameters are selected as required. Unlike Newton's method, all of them can be also applied in the case of a singular Hessian matrix. More importantly, by reduction, some of them avoid calculating the inverse of the Hessian matrix or an identical dimension matrix and some of them need only the diagonal elements of the Hessian matrix. In these cases, these algorithms still outperform the gradient descent method. The merits of the proposed optimization algorithm are illustrated by numerical experiments.

♻ ☆ Optimization Methods Rooting in Optimal Control

Huanshui Zhang, Hongxia Wang

In the paper, we propose solving optimization problems (OPs) and understanding the Newton method from the optimal control view. We propose a new optimization algorithm based on the optimal control problem (OCP). The algorithm features converging more rapidly than gradient descent, meanwhile, it is superior to Newton's method because it is not divergent in general and can be applied in the case of a singular Hessian matrix. These merits are supported by the convergence analysis for the algorithm in the paper. We also point out that the convergence rate of the proposed algorithm is inversely proportional to the magnitude of the control weight matrix and proportional to the control terminal time inherited from OCP.

♻ ☆ Optimal Control of Discrete-Time Nonlinear Systems

Chuanzhi Lv, Xunmin Yin, Hongdan Li, Huanshui Zhang

This paper focuses on optimal control problem for a class of discrete-time nonlinear systems. In practical applications, computation time is a crucial consideration when solving nonlinear optimal control problems, especially under real-time constraints. While linearization methods are computationally efficient, their inherent low accuracy can compromise control precision and overall performance. To address this challenge, this study proposes a novel approach based on the optimal control method. Firstly, the original optimal control problem is transformed into an equivalent optimization problem, which is resolved using the Pontryagin's maximum principle, and a superlinear convergence algorithm is presented. Furthermore, to improve computation efficiency, explicit formulas for computing both the gradient and hessian matrix of the cost function are proposed. Finally, the effectiveness of the proposed algorithm is validated through simulations and experiments on a linear quadratic regulator problem and an automatic guided vehicle trajectory tracking problem, demonstrating its ability for real-time online precise control.

♻ ☆ Set-point control and local stability for flat nonlinear systems using model-following control

Julian Willkomm, Kai Wulff, Johann Reger

We consider the set-point control problem for nonlinear systems with flat output that are subject to perturbations. The nonlinear dynamics as well as the perturbations are locally Lipschitz. We apply the model-following control (MFC) approach which consists of a model control loop (MCL) for a feedforward generation and a process control loop (PCL) that compensates the perturbations using high-gain feedback. We analyse the resulting closed-loop system and discuss its relation to a standard flatness-based high-gain approach. In particular we analyse the estimated region of attraction provided by a quadratic Lyapunov function. A case study illustrates the approach and quantifies the region of attraction obtained for each control approach. Using the initial condition of the model control loop as tuning parameter for the MFC design, provides that a significantly larger region of attraction can be guaranteed compared to a conventional single-loop high-gain design.

♻ ☆ Decentralized Bilevel Optimization: A Perspective from Transient Iteration Complexity

Boao Kong, Shuchen Zhu, Songtao Lu, Xinmeng Huang, Kun Yuan

Stochastic bilevel optimization (SBO) is becoming increasingly essential in machine learning due to its versatility in handling nested structures. To address large-scale SBO, decentralized approaches have emerged as effective paradigms in which nodes communicate with immediate neighbors without a central server, thereby improving communication efficiency and enhancing algorithmic robustness. However, most decentralized SBO algorithms focus solely on asymptotic convergence rates, overlooking transient iteration complexity-the number of iterations required before asymptotic rates dominate, which results in limited understanding of the influence of network topology, data heterogeneity, and the nested bilevel algorithmic structures. To address this issue, this paper introduces D-SOBA, a Decentralized Stochastic One-loop Bilevel Algorithm framework. D-SOBA comprises two variants: D-SOBA-SO, which incorporates second-order Hessian and Jacobian matrices, and D-SOBA-FO, which relies entirely on first-order gradients. We provide a comprehensive non-asymptotic convergence analysis and establish the transient iteration complexity of D-SOBA. This provides the first theoretical understanding of how network topology, data heterogeneity, and nested bilevel structures influence decentralized SBO. Extensive experimental results demonstrate the efficiency and theoretical advantages of D-SOBA.

comment: 59 pages, 7 figures

☆ Easi3R: Estimating Disentangled Motion from DUSt3R Without Training

Xingyu Chen, Yue Chen, Yuliang Xiu, Andreas Geiger, Anpei Chen

Recent advances in DUSt3R have enabled robust estimation of dense point clouds and camera parameters of static scenes, leveraging Transformer network architectures and direct supervision on large-scale 3D datasets. In contrast, the limited scale and diversity of available 4D datasets present a major bottleneck for training a highly generalizable 4D model. This constraint has driven conventional 4D methods to fine-tune 3D models on scalable dynamic video data with additional geometric priors such as optical flow and depths. In this work, we take an opposite path and introduce Easi3R, a simple yet efficient training-free method for 4D reconstruction. Our approach applies attention adaptation during inference, eliminating the need for from-scratch pre-training or network fine-tuning. We find that the attention layers in DUSt3R inherently encode rich information about camera and object motion. By carefully disentangling these attention maps, we achieve accurate dynamic region segmentation, camera pose estimation, and 4D dense point map reconstruction. Extensive experiments on real-world dynamic videos demonstrate that our lightweight attention adaptation significantly outperforms previous state-of-the-art methods that are trained or finetuned on extensive dynamic datasets. Our code is publicly available for research purpose at https://easi3r.github.io/

comment: Page: https://easi3r.github.io/ Code: https://github.com/Inception3D/Easi3R

☆ SU-YOLO: Spiking Neural Network for Efficient Underwater Object Detection

Chenyang Li, Wenxuan Liu, Guoqiang Gong, Xiaobo Ding, Xian Zhong

Underwater object detection is critical for oceanic research and industrial safety inspections. However, the complex optical environment and the limited resources of underwater equipment pose significant challenges to achieving high accuracy and low power consumption. To address these issues, we propose Spiking Underwater YOLO (SU-YOLO), a Spiking Neural Network (SNN) model. Leveraging the lightweight and energy-efficient properties of SNNs, SU-YOLO incorporates a novel spike-based underwater image denoising method based solely on integer addition, which enhances the quality of feature maps with minimal computational overhead. In addition, we introduce Separated Batch Normalization (SeBN), a technique that normalizes feature maps independently across multiple time steps and is optimized for integration with residual structures to capture the temporal dynamics of SNNs more effectively. The redesigned spiking residual blocks integrate the Cross Stage Partial Network (CSPNet) with the YOLO architecture to mitigate spike degradation and enhance the model's feature extraction capabilities. Experimental results on URPC2019 underwater dataset demonstrate that SU-YOLO achieves mAP of 78.8% with 6.97M parameters and an energy consumption of 2.98 mJ, surpassing mainstream SNN models in both detection accuracy and computational efficiency. These results underscore the potential of SNNs for engineering applications. The code is available in https://github.com/lwxfight/snn-underwater.

☆ Consistent Subject Generation via Contrastive Instantiated Concepts

Lee Hsin-Ying, Kelvin C. K. Chan, Ming-Hsuan Yang

While text-to-image generative models can synthesize diverse and faithful contents, subject variation across multiple creations limits the application in long content generation. Existing approaches require time-consuming tuning, references for all subjects, or access to other creations. We introduce Contrastive Concept Instantiation (CoCoIns) to effectively synthesize consistent subjects across multiple independent creations. The framework consists of a generative model and a mapping network, which transforms input latent codes into pseudo-words associated with certain instances of concepts. Users can generate consistent subjects with the same latent codes. To construct such associations, we propose a contrastive learning approach that trains the network to differentiate the combination of prompts and latent codes. Extensive evaluations of human faces with a single subject show that CoCoIns performs comparably to existing methods while maintaining higher flexibility. We also demonstrate the potential of extending CoCoIns to multiple subjects and other object categories.

comment: Project page: https://contrastive-concept-instantiation.github.io

☆ Free360: Layered Gaussian Splatting for Unbounded 360-Degree View Synthesis from Extremely Sparse and Unposed Views CVPR 2025

Chong Bao, Xiyu Zhang, Zehao Yu, Jiale Shi, Guofeng Zhang, Songyou Peng, Zhaopeng Cui

Neural rendering has demonstrated remarkable success in high-quality 3D neural reconstruction and novel view synthesis with dense input views and accurate poses. However, applying it to extremely sparse, unposed views in unbounded 360{\deg} scenes remains a challenging problem. In this paper, we propose a novel neural rendering framework to accomplish the unposed and extremely sparse-view 3D reconstruction in unbounded 360{\deg} scenes. To resolve the spatial ambiguity inherent in unbounded scenes with sparse input views, we propose a layered Gaussian-based representation to effectively model the scene with distinct spatial layers. By employing a dense stereo reconstruction model to recover coarse geometry, we introduce a layer-specific bootstrap optimization to refine the noise and fill occluded regions in the reconstruction. Furthermore, we propose an iterative fusion of reconstruction and generation alongside an uncertainty-aware training approach to facilitate mutual conditioning and enhancement between these two processes. Comprehensive experiments show that our approach outperforms existing state-of-the-art methods in terms of rendering quality and surface reconstruction accuracy. Project page: https://zju3dv.github.io/free360/

comment: Accepted to CVPR 2025. Project Page: https://zju3dv.github.io/free360/

☆ UniOcc: A Unified Benchmark for Occupancy Forecasting and Prediction in Autonomous Driving

Yuping Wang, Xiangyu Huang, Xiaokang Sun, Mingxuan Yan, Shuo Xing, Zhengzhong Tu, Jiachen Li

We introduce UniOcc, a comprehensive, unified benchmark for occupancy forecasting (i.e., predicting future occupancies based on historical information) and current-frame occupancy prediction from camera images. UniOcc unifies data from multiple real-world datasets (i.e., nuScenes, Waymo) and high-fidelity driving simulators (i.e., CARLA, OpenCOOD), which provides 2D/3D occupancy labels with per-voxel flow annotations and support for cooperative autonomous driving. In terms of evaluation, unlike existing studies that rely on suboptimal pseudo labels for evaluation, UniOcc incorporates novel metrics that do not depend on ground-truth occupancy, enabling robust assessment of additional aspects of occupancy quality. Through extensive experiments on state-of-the-art models, we demonstrate that large-scale, diverse training data and explicit flow information significantly enhance occupancy prediction and forecasting performance.

comment: 14 pages; Dataset: https://huggingface.co/datasets/tasl-lab/uniocc; Code: https://github.com/tasl-lab/UniOcc

☆ Any2Caption:Interpreting Any Condition to Caption for Controllable Video Generation

Shengqiong Wu, Weicai Ye, Jiahao Wang, Quande Liu, Xintao Wang, Pengfei Wan, Di Zhang, Kun Gai, Shuicheng Yan, Hao Fei, Tat-Seng Chua

To address the bottleneck of accurate user intent interpretation within the current video generation community, we present Any2Caption, a novel framework for controllable video generation under any condition. The key idea is to decouple various condition interpretation steps from the video synthesis step. By leveraging modern multimodal large language models (MLLMs), Any2Caption interprets diverse inputs--text, images, videos, and specialized cues such as region, motion, and camera poses--into dense, structured captions that offer backbone video generators with better guidance. We also introduce Any2CapIns, a large-scale dataset with 337K instances and 407K conditions for any-condition-to-caption instruction tuning. Comprehensive evaluations demonstrate significant improvements of our system in controllability and video quality across various aspects of existing video generation models. Project Page: https://sqwu.top/Any2Cap/

comment: Project Page: https://sqwu.top/Any2Cap/

☆ Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1

Yi Chen, Yuying Ge, Rui Wang, Yixiao Ge, Lu Qiu, Ying Shan, Xihui Liu

Recent advancements in Chain of Thought (COT) generation have significantly improved the reasoning capabilities of Large Language Models (LLMs), with reinforcement learning (RL) emerging as an effective post-training approach. Multimodal Large Language Models (MLLMs) inherit this reasoning potential but remain underexplored in tasks requiring both perception and logical reasoning. To address this, we introduce SEED-Bench-R1, a benchmark designed to systematically evaluate post-training methods for MLLMs in video understanding. It includes intricate real-world videos and complex everyday planning tasks in the format of multiple-choice questions, requiring sophisticated perception and reasoning. SEED-Bench-R1 assesses generalization through a three-level hierarchy: in-distribution, cross-environment, and cross-environment-task scenarios, equipped with a large-scale training dataset with easily verifiable ground-truth answers. Using Qwen2-VL-Instruct-7B as a base model, we compare RL with supervised fine-tuning (SFT), demonstrating RL's data efficiency and superior performance on both in-distribution and out-of-distribution tasks, even outperforming SFT on general video understanding benchmarks like LongVideoBench. Our detailed analysis reveals that RL enhances visual perception but often produces less logically coherent reasoning chains. We identify key limitations such as inconsistent reasoning and overlooked visual cues, and suggest future improvements in base model reasoning, reward modeling, and RL robustness against noisy signals.

comment: Technical Report (In Progress); Code released at: https://github.com/TencentARC/SEED-Bench-R1

☆ ERUPT: Efficient Rendering with Unposed Patch Transformer CVPR 2025

Maxim V. Shugaev, Vincent Chen, Maxim Karrenbach, Kyle Ashley, Bridget Kennedy, Naresh P. Cuntoor

This work addresses the problem of novel view synthesis in diverse scenes from small collections of RGB images. We propose ERUPT (Efficient Rendering with Unposed Patch Transformer) a state-of-the-art scene reconstruction model capable of efficient scene rendering using unposed imagery. We introduce patch-based querying, in contrast to existing pixel-based queries, to reduce the compute required to render a target view. This makes our model highly efficient both during training and at inference, capable of rendering at 600 fps on commercial hardware. Notably, our model is designed to use a learned latent camera pose which allows for training using unposed targets in datasets with sparse or inaccurate ground truth camera pose. We show that our approach can generalize on large real-world data and introduce a new benchmark dataset (MSVS-1M) for latent view synthesis using street-view imagery collected from Mapillary. In contrast to NeRF and Gaussian Splatting, which require dense imagery and precise metadata, ERUPT can render novel views of arbitrary scenes with as few as five unposed input images. ERUPT achieves better rendered image quality than current state-of-the-art methods for unposed image synthesis tasks, reduces labeled data requirements by ~95\% and decreases computational requirements by an order of magnitude, providing efficient novel view synthesis for diverse real-world scenes.

comment: Accepted to CVPR 2025

☆ Adapting Vision Foundation Models for Real-time Ultrasound Image Segmentation

Xiaoran Zhang, Eric Z. Chen, Lin Zhao, Xiao Chen, Yikang Liu, Boris Maihe, James S. Duncan, Terrence Chen, Shanhui Sun

We propose a novel approach that adapts hierarchical vision foundation models for real-time ultrasound image segmentation. Existing ultrasound segmentation methods often struggle with adaptability to new tasks, relying on costly manual annotations, while real-time approaches generally fail to match state-of-the-art performance. To overcome these limitations, we introduce an adaptive framework that leverages the vision foundation model Hiera to extract multi-scale features, interleaved with DINOv2 representations to enhance visual expressiveness. These enriched features are then decoded to produce precise and robust segmentation. We conduct extensive evaluations on six public datasets and one in-house dataset, covering both cardiac and thyroid ultrasound segmentation. Experiments show that our approach outperforms state-of-the-art methods across multiple datasets and excels with limited supervision, surpassing nnUNet by over 20\% on average in the 1\% and 10\% data settings. Our method achieves $\sim$77 FPS inference speed with TensorRT on a single GPU, enabling real-time clinical applications.

☆ StochasticSplats: Stochastic Rasterization for Sorting-Free 3D Gaussian Splatting

Shakiba Kheradmand, Delio Vicini, George Kopanas, Dmitry Lagun, Kwang Moo Yi, Mark Matthews, Andrea Tagliasacchi

3D Gaussian splatting (3DGS) is a popular radiance field method, with many application-specific extensions. Most variants rely on the same core algorithm: depth-sorting of Gaussian splats then rasterizing in primitive order. This ensures correct alpha compositing, but can cause rendering artifacts due to built-in approximations. Moreover, for a fixed representation, sorted rendering offers little control over render cost and visual fidelity. For example, and counter-intuitively, rendering a lower-resolution image is not necessarily faster. In this work, we address the above limitations by combining 3D Gaussian splatting with stochastic rasterization. Concretely, we leverage an unbiased Monte Carlo estimator of the volume rendering equation. This removes the need for sorting, and allows for accurate 3D blending of overlapping Gaussians. The number of Monte Carlo samples further imbues 3DGS with a way to trade off computation time and quality. We implement our method using OpenGL shaders, enabling efficient rendering on modern GPU hardware. At a reasonable visual quality, our method renders more than four times faster than sorted rasterization.

☆ InstructRestore: Region-Customized Image Restoration with Human Instructions

Shuaizheng Liu, Jianqi Ma, Lingchen Sun, Xiangtao Kong, Lei Zhang

Despite the significant progress in diffusion prior-based image restoration, most existing methods apply uniform processing to the entire image, lacking the capability to perform region-customized image restoration according to user instructions. In this work, we propose a new framework, namely InstructRestore, to perform region-adjustable image restoration following human instructions. To achieve this, we first develop a data generation engine to produce training triplets, each consisting of a high-quality image, the target region description, and the corresponding region mask. With this engine and careful data screening, we construct a comprehensive dataset comprising 536,945 triplets to support the training and evaluation of this task. We then examine how to integrate the low-quality image features under the ControlNet architecture to adjust the degree of image details enhancement. Consequently, we develop a ControlNet-like model to identify the target region and allocate different integration scales to the target and surrounding regions, enabling region-customized image restoration that aligns with user instructions. Experimental results demonstrate that our proposed InstructRestore approach enables effective human-instructed image restoration, such as images with bokeh effects and user-instructed local enhancement. Our work advances the investigation of interactive image restoration and enhancement techniques. Data, code, and models will be found at https://github.com/shuaizhengliu/InstructRestore.git.

☆ ORAL: Prompting Your Large-Scale LoRAs via Conditional Recurrent Diffusion

Rana Muhammad Shahroz Khan, Dongwen Tang, Pingzhi Li, Kai Wang, Tianlong Chen

Parameter generation has emerged as a novel paradigm for neural network development, offering an alternative to traditional neural network training by synthesizing high-quality model weights directly. In the context of Low-Rank Adaptation (LoRA) for evolving ($\textit{i.e.}$, constantly updated) large language models (LLMs), this approach promises efficient adaptation without costly retraining. However, existing methods face critical limitations in simultaneously achieving scalability and controllability. In this paper, we introduce $\texttt{ORAL}$, a novel $\textbf{conditional recurrent diffusion}$ framework that addresses these challenges. $\texttt{ORAL}$ incorporates a novel conditioning mechanism that integrates model architecture and textual task specifications, enabling the generation of task-specific LoRA parameters that can seamlessly transfer across evolving foundation models. Our approach successfully scales to billions-of-parameter LLMs and maintains controllability. Through extensive experiments across seven language tasks, four vision tasks, and three multimodal tasks using five pre-trained LLMs, we demonstrate that $\texttt{ORAL}$ generates high-quality LoRA parameters that achieve comparable or superior performance to vanilla trained counterparts.

☆ PathOrchestra: A Comprehensive Foundation Model for Computational Pathology with Over 100 Diverse Clinical-Grade Tasks

Fang Yan, Jianfeng Wu, Jiawen Li, Wei Wang, Jiaxuan Lu, Wen Chen, Zizhao Gao, Jianan Li, Hong Yan, Jiabo Ma, Minda Chen, Yang Lu, Qing Chen, Yizhi Wang, Xitong Ling, Xuenian Wang, Zihan Wang, Qiang Huang, Shengyi Hua, Mianxin Liu, Lei Ma, Tian Shen, Xiaofan Zhang, Yonghong He, Hao Chen, Shaoting Zhang, Zhe Wang

The complexity and variability inherent in high-resolution pathological images present significant challenges in computational pathology. While pathology foundation models leveraging AI have catalyzed transformative advancements, their development demands large-scale datasets, considerable storage capacity, and substantial computational resources. Furthermore, ensuring their clinical applicability and generalizability requires rigorous validation across a broad spectrum of clinical tasks. Here, we present PathOrchestra, a versatile pathology foundation model trained via self-supervised learning on a dataset comprising 300K pathological slides from 20 tissue and organ types across multiple centers. The model was rigorously evaluated on 112 clinical tasks using a combination of 61 private and 51 public datasets. These tasks encompass digital slide preprocessing, pan-cancer classification, lesion identification, multi-cancer subtype classification, biomarker assessment, gene expression prediction, and the generation of structured reports. PathOrchestra demonstrated exceptional performance across 27,755 WSIs and 9,415,729 ROIs, achieving over 0.950 accuracy in 47 tasks, including pan-cancer classification across various organs, lymphoma subtype diagnosis, and bladder cancer screening. Notably, it is the first model to generate structured reports for high-incidence colorectal cancer and diagnostically complex lymphoma-areas that are infrequently addressed by foundational models but hold immense clinical potential. Overall, PathOrchestra exemplifies the feasibility and efficacy of a large-scale, self-supervised pathology foundation model, validated across a broad range of clinical-grade tasks. Its high accuracy and reduced reliance on extensive data annotation underline its potential for clinical integration, offering a pathway toward more efficient and high-quality medical services.

☆ Self-Supervised Pretraining for Aerial Road Extraction

Rupert Polley, Sai Vignesh Abishek Deenadayalan, J. Marius Zöllner

Deep neural networks for aerial image segmentation require large amounts of labeled data, but high-quality aerial datasets with precise annotations are scarce and costly to produce. To address this limitation, we propose a self-supervised pretraining method that improves segmentation performance while reducing reliance on labeled data. Our approach uses inpainting-based pretraining, where the model learns to reconstruct missing regions in aerial images, capturing their inherent structure before being fine-tuned for road extraction. This method improves generalization, enhances robustness to domain shifts, and is invariant to model architecture and dataset choice. Experiments show that our pretraining significantly boosts segmentation accuracy, especially in low-data regimes, making it a scalable solution for aerial image analysis.

☆ Can Test-Time Scaling Improve World Foundation Model?

Wenyan Cong, Hanqing Zhu, Peihao Wang, Bangya Liu, Dejia Xu, Kevin Wang, David Z. Pan, Yan Wang, Zhiwen Fan, Zhangyang Wang

World foundation models, which simulate the physical world by predicting future states from current observations and inputs, have become central to many applications in physical intelligence, including autonomous driving and robotics. However, these models require substantial computational resources for pretraining and are further constrained by available data during post-training. As such, scaling computation at test time emerges as both a critical and practical alternative to traditional model enlargement or re-training. In this work, we introduce SWIFT, a test-time scaling framework tailored for WFMs. SWIFT integrates our extensible WFM evaluation toolkit with process-level inference strategies, including fast tokenization, probability-based Top-K pruning, and efficient beam search. Empirical results on the COSMOS model demonstrate that test-time scaling exists even in a compute-optimal way. Our findings reveal that test-time scaling laws hold for WFMs and that SWIFT provides a scalable and effective pathway for improving WFM inference without retraining or increasing model size. The code is available at https://github.com/Mia-Cong/SWIFT.git.

☆ Point Tracking in Surgery--The 2024 Surgical Tattoos in Infrared (STIR) Challenge

Adam Schmidt, Mert Asim Karaoglu, Soham Sinha, Mingang Jang, Ho-Gun Ha, Kyungmin Jung, Kyeongmo Gu, Ihsan Ullah, Hyunki Lee, Jonáš Šerých, Michal Neoral, Jiří Matas, Rulin Zhou, Wenlong He, An Wang, Hongliang Ren, Bruno Silva, Sandro Queirós, Estêvão Lima, João L. Vilaça, Shunsuke Kikuchi, Atsushi Kouno, Hiroki Matsuzaki, Tongtong Li, Yulu Chen, Ling Li, Xiang Ma, Xiaojian Li, Mona Sheikh Zeinoddin, Xu Wang, Zafer Tandogdu, Greg Shaw, Evangelos Mazomenos, Danail Stoyanov, Yuxin Chen, Zijian Wu, Alexander Ladikos, Simon DiMaio, Septimiu E. Salcudean, Omid Mohareri

Understanding tissue motion in surgery is crucial to enable applications in downstream tasks such as segmentation, 3D reconstruction, virtual tissue landmarking, autonomous probe-based scanning, and subtask autonomy. Labeled data are essential to enabling algorithms in these downstream tasks since they allow us to quantify and train algorithms. This paper introduces a point tracking challenge to address this, wherein participants can submit their algorithms for quantification. The submitted algorithms are evaluated using a dataset named surgical tattoos in infrared (STIR), with the challenge aptly named the STIR Challenge 2024. The STIR Challenge 2024 comprises two quantitative components: accuracy and efficiency. The accuracy component tests the accuracy of algorithms on in vivo and ex vivo sequences. The efficiency component tests the latency of algorithm inference. The challenge was conducted as a part of MICCAI EndoVis 2024. In this challenge, we had 8 total teams, with 4 teams submitting before and 4 submitting after challenge day. This paper details the STIR Challenge 2024, which serves to move the field towards more accurate and efficient algorithms for spatial understanding in surgery. In this paper we summarize the design, submissions, and results from the challenge. The challenge dataset is available here: https://zenodo.org/records/14803158 , and the code for baseline models and metric calculation is available here: https://github.com/athaddius/STIRMetrics

☆ Order Matters: On Parameter-Efficient Image-to-Video Probing for Recognizing Nearly Symmetric Actions

Thinesh Thiyakesan Ponbagavathi, Alina Roitberg

We study parameter-efficient image-to-video probing for the unaddressed challenge of recognizing nearly symmetric actions - visually similar actions that unfold in opposite temporal order (e.g., opening vs. closing a bottle). Existing probing mechanisms for image-pretrained models, such as DinoV2 and CLIP, rely on attention mechanism for temporal modeling but are inherently permutation-invariant, leading to identical predictions regardless of frame order. To address this, we introduce Self-attentive Temporal Embedding Probing (STEP), a simple yet effective approach designed to enforce temporal sensitivity in parameter-efficient image-to-video transfer. STEP enhances self-attentive probing with three key modifications: (1) a learnable frame-wise positional encoding, explicitly encoding temporal order; (2) a single global CLS token, for sequence coherence; and (3) a simplified attention mechanism to improve parameter efficiency. STEP outperforms existing image-to-video probing mechanisms by 3-15% across four activity recognition benchmarks with only 1/3 of the learnable parameters. On two datasets, it surpasses all published methods, including fully fine-tuned models. STEP shows a distinct advantage in recognizing nearly symmetric actions, surpassing other probing mechanisms by 9-19%. and parameter-heavier PEFT-based transfer methods by 5-15%. Code and models will be made publicly available.

☆ Style Quantization for Data-Efficient GAN Training

Jian Wang, Xin Lan, Jizhe Zhou, Yuxin Tian, Jiancheng Lv

Under limited data setting, GANs often struggle to navigate and effectively exploit the input latent space. Consequently, images generated from adjacent variables in a sparse input latent space may exhibit significant discrepancies in realism, leading to suboptimal consistency regularization (CR) outcomes. To address this, we propose \textit{SQ-GAN}, a novel approach that enhances CR by introducing a style space quantization scheme. This method transforms the sparse, continuous input latent space into a compact, structured discrete proxy space, allowing each element to correspond to a specific real data point, thereby improving CR performance. Instead of direct quantization, we first map the input latent variables into a less entangled ``style'' space and apply quantization using a learnable codebook. This enables each quantized code to control distinct factors of variation. Additionally, we optimize the optimal transport distance to align the codebook codes with features extracted from the training data by a foundation model, embedding external knowledge into the codebook and establishing a semantically rich vocabulary that properly describes the training dataset. Extensive experiments demonstrate significant improvements in both discriminator robustness and generation quality with our method.

☆ Learning Velocity and Acceleration: Self-Supervised Motion Consistency for Pedestrian Trajectory Prediction

Yizhou Huang, Yihua Cheng, Kezhi Wang

Understanding human motion is crucial for accurate pedestrian trajectory prediction. Conventional methods typically rely on supervised learning, where ground-truth labels are directly optimized against predicted trajectories. This amplifies the limitations caused by long-tailed data distributions, making it difficult for the model to capture abnormal behaviors. In this work, we propose a self-supervised pedestrian trajectory prediction framework that explicitly models position, velocity, and acceleration. We leverage velocity and acceleration information to enhance position prediction through feature injection and a self-supervised motion consistency mechanism. Our model hierarchically injects velocity features into the position stream. Acceleration features are injected into the velocity stream. This enables the model to predict position, velocity, and acceleration jointly. From the predicted position, we compute corresponding pseudo velocity and acceleration, allowing the model to learn from data-generated pseudo labels and thus achieve self-supervised learning. We further design a motion consistency evaluation strategy grounded in physical principles; it selects the most reasonable predicted motion trend by comparing it with historical dynamics and uses this trend to guide and constrain trajectory generation. We conduct experiments on the ETH-UCY and Stanford Drone datasets, demonstrating that our method achieves state-of-the-art performance on both datasets.

☆ Visual Acoustic Fields

Yuelei Li, Hyunjin Kim, Fangneng Zhan, Ri-Zhao Qiu, Mazeyu Ji, Xiaojun Shan, Xueyan Zou, Paul Liang, Hanspeter Pfister, Xiaolong Wang

Objects produce different sounds when hit, and humans can intuitively infer how an object might sound based on its appearance and material properties. Inspired by this intuition, we propose Visual Acoustic Fields, a framework that bridges hitting sounds and visual signals within a 3D space using 3D Gaussian Splatting (3DGS). Our approach features two key modules: sound generation and sound localization. The sound generation module leverages a conditional diffusion model, which takes multiscale features rendered from a feature-augmented 3DGS to generate realistic hitting sounds. Meanwhile, the sound localization module enables querying the 3D scene, represented by the feature-augmented 3DGS, to localize hitting positions based on the sound sources. To support this framework, we introduce a novel pipeline for collecting scene-level visual-sound sample pairs, achieving alignment between captured images, impact locations, and corresponding sounds. To the best of our knowledge, this is the first dataset to connect visual and acoustic signals in a 3D context. Extensive experiments on our dataset demonstrate the effectiveness of Visual Acoustic Fields in generating plausible impact sounds and accurately localizing impact sources. Our project page is at https://yuelei0428.github.io/projects/Visual-Acoustic-Fields/.

☆ FakeScope: Large Multimodal Expert Model for Transparent AI-Generated Image Forensics

Yixuan Li, Yu Tian, Yipo Huang, Wei Lu, Shiqi Wang, Weisi Lin, Anderson Rocha

The rapid and unrestrained advancement of generative artificial intelligence (AI) presents a double-edged sword: while enabling unprecedented creativity, it also facilitates the generation of highly convincing deceptive content, undermining societal trust. As image generation techniques become increasingly sophisticated, detecting synthetic images is no longer just a binary task: it necessitates interpretable, context-aware methodologies that enhance trustworthiness and transparency. However, existing detection models primarily focus on classification, offering limited explanatory insights into image authenticity. In this work, we propose FakeScope, an expert multimodal model (LMM) tailored for AI-generated image forensics, which not only identifies AI-synthetic images with high accuracy but also provides rich, interpretable, and query-driven forensic insights. We first construct FakeChain dataset that contains linguistic authenticity reasoning based on visual trace evidence, developed through a novel human-machine collaborative framework. Building upon it, we further present FakeInstruct, the largest multimodal instruction tuning dataset containing 2 million visual instructions tailored to enhance forensic awareness in LMMs. FakeScope achieves state-of-the-art performance in both closed-ended and open-ended forensic scenarios. It can distinguish synthetic images with high accuracy while offering coherent and insightful explanations, free-form discussions on fine-grained forgery attributes, and actionable enhancement strategies. Notably, despite being trained exclusively on qualitative hard labels, FakeScope demonstrates remarkable zero-shot quantitative capability on detection, enabled by our proposed token-based probability estimation strategy. Furthermore, FakeScope exhibits strong generalization and in-the-wild ability, ensuring its applicability in real-world scenarios.

☆ Beyond a Single Mode: GAN Ensembles for Diverse Medical Data Generation

Lorenzo Tronchin, Tommy Löfstedt, Paolo Soda, Valerio Guarrasi

The advancement of generative AI, particularly in medical imaging, confronts the trilemma of ensuring high fidelity, diversity, and efficiency in synthetic data generation. While Generative Adversarial Networks (GANs) have shown promise across various applications, they still face challenges like mode collapse and insufficient coverage of real data distributions. This work explores the use of GAN ensembles to overcome these limitations, specifically in the context of medical imaging. By solving a multi-objective optimisation problem that balances fidelity and diversity, we propose a method for selecting an optimal ensemble of GANs tailored for medical data. The selected ensemble is capable of generating diverse synthetic medical images that are representative of true data distributions and computationally efficient. Each model in the ensemble brings a unique contribution, ensuring minimal redundancy. We conducted a comprehensive evaluation using three distinct medical datasets, testing 22 different GAN architectures with various loss functions and regularisation techniques. By sampling models at different training epochs, we crafted 110 unique configurations. The results highlight the capability of GAN ensembles to enhance the quality and utility of synthetic medical images, thereby improving the efficacy of downstream tasks such as diagnostic modelling.

☆ Pre-training with 3D Synthetic Data: Learning 3D Point Cloud Instance Segmentation from 3D Synthetic Scenes

Daichi Otsuka, Shinichi Mae, Ryosuke Yamada, Hirokatsu Kataoka

In the recent years, the research community has witnessed growing use of 3D point cloud data for the high applicability in various real-world applications. By means of 3D point cloud, this modality enables to consider the actual size and spatial understanding. The applied fields include mechanical control of robots, vehicles, or other real-world systems. Along this line, we would like to improve 3D point cloud instance segmentation which has emerged as a particularly promising approach for these applications. However, the creation of 3D point cloud datasets entails enormous costs compared to 2D image datasets. To train a model of 3D point cloud instance segmentation, it is necessary not only to assign categories but also to provide detailed annotations for each point in the large-scale 3D space. Meanwhile, the increase of recent proposals for generative models in 3D domain has spurred proposals for using a generative model to create 3D point cloud data. In this work, we propose a pre-training with 3D synthetic data to train a 3D point cloud instance segmentation model based on generative model for 3D scenes represented by point cloud data. We directly generate 3D point cloud data with Point-E for inserting a generated data into a 3D scene. More recently in 2025, although there are other accurate 3D generation models, even using the Point-E as an early 3D generative model can effectively support the pre-training with 3D synthetic data. In the experimental section, we compare our pre-training method with baseline methods indicated improved performance, demonstrating the efficacy of 3D generative models for 3D point cloud instance segmentation.

☆ MB-ORES: A Multi-Branch Object Reasoner for Visual Grounding in Remote Sensing

Karim Radouane, Hanane Azzag, Mustapha lebbah

We propose a unified framework that integrates object detection (OD) and visual grounding (VG) for remote sensing (RS) imagery. To support conventional OD and establish an intuitive prior for VG task, we fine-tune an open-set object detector using referring expression data, framing it as a partially supervised OD task. In the first stage, we construct a graph representation of each image, comprising object queries, class embeddings, and proposal locations. Then, our task-aware architecture processes this graph to perform the VG task. The model consists of: (i) a multi-branch network that integrates spatial, visual, and categorical features to generate task-aware proposals, and (ii) an object reasoning network that assigns probabilities across proposals, followed by a soft selection mechanism for final referring object localization. Our model demonstrates superior performance on the OPT-RSVG and DIOR-RSVG datasets, achieving significant improvements over state-of-the-art methods while retaining classical OD capabilities. The code will be available in our repository: \url{https://github.com/rd20karim/MB-ORES}.

☆ DiET-GS: Diffusion Prior and Event Stream-Assisted Motion Deblurring 3D Gaussian Splatting CVPR 2025

Seungjun Lee, Gim Hee Lee

Reconstructing sharp 3D representations from blurry multi-view images are long-standing problem in computer vision. Recent works attempt to enhance high-quality novel view synthesis from the motion blur by leveraging event-based cameras, benefiting from high dynamic range and microsecond temporal resolution. However, they often reach sub-optimal visual quality in either restoring inaccurate color or losing fine-grained details. In this paper, we present DiET-GS, a diffusion prior and event stream-assisted motion deblurring 3DGS. Our framework effectively leverages both blur-free event streams and diffusion prior in a two-stage training strategy. Specifically, we introduce the novel framework to constraint 3DGS with event double integral, achieving both accurate color and well-defined details. Additionally, we propose a simple technique to leverage diffusion prior to further enhance the edge details. Qualitative and quantitative results on both synthetic and real-world data demonstrate that our DiET-GS is capable of producing significantly better quality of novel views compared to the existing baselines. Our project page is https://diet-gs.github.io

comment: CVPR 2025. Project Page: https://diet-gs.github.io

☆ CIBR: Cross-modal Information Bottleneck Regularization for Robust CLIP Generalization

Yingrui Ji, Xi Xiao, Gaofei Chen, Hao Xu, Chenrui Ma, Lijing Zhu, Aokun Liang, Jiansheng Chen

Contrastive Language-Image Pretraining (CLIP) has achieved remarkable success in cross-modal tasks such as zero-shot image classification and text-image retrieval by effectively aligning visual and textual representations. However, the theoretical foundations underlying CLIP's strong generalization remain unclear. In this work, we address this gap by proposing the Cross-modal Information Bottleneck (CIB) framework. CIB offers a principled interpretation of CLIP's contrastive learning objective as an implicit Information Bottleneck optimization. Under this view, the model maximizes shared cross-modal information while discarding modality-specific redundancies, thereby preserving essential semantic alignment across modalities. Building on this insight, we introduce a Cross-modal Information Bottleneck Regularization (CIBR) method that explicitly enforces these IB principles during training. CIBR introduces a penalty term to discourage modality-specific redundancy, thereby enhancing semantic alignment between image and text features. We validate CIBR on extensive vision-language benchmarks, including zero-shot classification across seven diverse image datasets and text-image retrieval on MSCOCO and Flickr30K. The results show consistent performance gains over standard CLIP. These findings provide the first theoretical understanding of CLIP's generalization through the IB lens. They also demonstrate practical improvements, offering guidance for future cross-modal representation learning.

☆ Navi-plus: Managing Ambiguous GUI Navigation Tasks with Follow-up

Ziming Cheng, Zhiyuan Huang, Junting Pan, Zhaohui Hou, Mingjie Zhan

Graphical user interfaces (GUI) automation agents are emerging as powerful tools, enabling humans to accomplish increasingly complex tasks on smart devices. However, users often inadvertently omit key information when conveying tasks, which hinders agent performance in the current agent paradigm that does not support immediate user intervention. To address this issue, we introduce a $\textbf{Self-Correction GUI Navigation}$ task that incorporates interactive information completion capabilities within GUI agents. We developed the $\textbf{Navi-plus}$ dataset with GUI follow-up question-answer pairs, alongside a $\textbf{Dual-Stream Trajectory Evaluation}$ method to benchmark this new capability. Our results show that agents equipped with the ability to ask GUI follow-up questions can fully recover their performance when faced with ambiguous user tasks.

☆ Foundation Models For Seismic Data Processing: An Extensive Review

Fabian Fuchs, Mario Ruben Fernandez, Norman Ettrich, Janis Keuper

Seismic processing plays a crucial role in transforming raw data into high-quality subsurface images, pivotal for various geoscience applications. Despite its importance, traditional seismic processing techniques face challenges such as noisy and damaged data and the reliance on manual, time-consuming workflows. The emergence of deep learning approaches has introduced effective and user-friendly alternatives, yet many of these deep learning approaches rely on synthetic datasets and specialized neural networks. Recently, foundation models have gained traction in the seismic domain, due to their success in natural imaging. This paper investigates the application of foundation models in seismic processing on the tasks: demultiple, interpolation, and denoising. It evaluates the impact of different model characteristics, such as pre-training technique and neural network architecture, on performance and efficiency. Rather than proposing a single seismic foundation model, this paper critically examines various natural image foundation models and suggest some promising candidates for future exploration.

☆ A Comparative Study of Scanpath Models in Graph-Based Visualization

Angela Lopez-Cardona, Parvin Emami, Sebastian Idesis, Saravanakumar Duraisamy, Luis A. Leiva, Ioannis Arapakis

Information Visualization (InfoVis) systems utilize visual representations to enhance data interpretation. Understanding how visual attention is allocated is essential for optimizing interface design. However, collecting Eye-tracking (ET) data presents challenges related to cost, privacy, and scalability. Computational models provide alternatives for predicting gaze patterns, thereby advancing InfoVis research. In our study, we conducted an ET experiment with 40 participants who analyzed graphs while responding to questions of varying complexity within the context of digital forensics. We compared human scanpaths with synthetic ones generated by models such as DeepGaze, UMSS, and Gazeformer. Our research evaluates the accuracy of these models and examines how question complexity and number of nodes influence performance. This work contributes to the development of predictive modeling in visual analytics, offering insights that can enhance the design and effectiveness of InfoVis systems.

☆ AI-Assisted Colonoscopy: Polyp Detection and Segmentation using Foundation Models

Uxue Delaquintana-Aramendi, Leire Benito-del-Valle, Aitor Alvarez-Gila, Javier Pascau, Luisa F Sánchez-Peralta, Artzai Picón, J Blas Pagador, Cristina L Saratxaga

In colonoscopy, 80% of the missed polyps could be detected with the help of Deep Learning models. In the search for algorithms capable of addressing this challenge, foundation models emerge as promising candidates. Their zero-shot or few-shot learning capabilities, facilitate generalization to new data or tasks without extensive fine-tuning. A concept that is particularly advantageous in the medical imaging domain, where large annotated datasets for traditional training are scarce. In this context, a comprehensive evaluation of foundation models for polyp segmentation was conducted, assessing both detection and delimitation. For the study, three different colonoscopy datasets have been employed to compare the performance of five different foundation models, DINOv2, YOLO-World, GroundingDINO, SAM and MedSAM, against two benchmark networks, YOLOv8 and Mask R-CNN. Results show that the success of foundation models in polyp characterization is highly dependent on domain specialization. For optimal performance in medical applications, domain-specific models are essential, and generic models require fine-tuning to achieve effective results. Through this specialization, foundation models demonstrated superior performance compared to state-of-the-art detection and segmentation models, with some models even excelling in zero-shot evaluation; outperforming fine-tuned models on unseen data.

comment: This work has been submitted to the IEEE TMI for possible publication

☆ PixelCAM: Pixel Class Activation Mapping for Histology Image Classification and ROI Localization

Alexis Guichemerre, Soufiane Belharbi, Mohammadhadi Shateri, Luke McCaffrey, Eric Granger

Weakly supervised object localization (WSOL) methods allow training models to classify images and localize ROIs. WSOL only requires low-cost image-class annotations yet provides a visually interpretable classifier, which is important in histology image analysis. Standard WSOL methods rely on class activation mapping (CAM) methods to produce spatial localization maps according to a single- or two-step strategy. While both strategies have made significant progress, they still face several limitations with histology images. Single-step methods can easily result in under- or over-activation due to the limited visual ROI saliency in histology images and the limited localization cues. They also face the well-known issue of asynchronous convergence between classification and localization tasks. The two-step approach is sub-optimal because it is tied to a frozen classifier, limiting the capacity for localization. Moreover, these methods also struggle when applied to out-of-distribution (OOD) datasets. In this paper, a multi-task approach for WSOL is introduced for simultaneous training of both tasks to address the asynchronous convergence problem. In particular, localization is performed in the pixel-feature space of an image encoder that is shared with classification. This allows learning discriminant features and accurate delineation of foreground/background regions to support ROI localization and image classification. We propose PixelCAM, a cost-effective foreground/background pixel-wise classifier in the pixel-feature space that allows for spatial object localization. PixelCAM is trained using pixel pseudo-labels collected from a pretrained WSOL model. Both image and pixel-wise classifiers are trained simultaneously using standard gradient descent. In addition, our pixel classifier can easily be integrated into CNN- and transformer-based architectures without any modifications.

comment: 32 pages, 20 figures, Medical Imaging with Deep Learning (MIDL 2025)

☆ It's a (Blind) Match! Towards Vision-Language Correspondence without Parallel Data CVPR 2025

Dominik Schnaus, Nikita Araslanov, Daniel Cremers

The platonic representation hypothesis suggests that vision and language embeddings become more homogeneous as model and dataset sizes increase. In particular, pairwise distances within each modality become more similar. This suggests that as foundation models mature, it may become possible to match vision and language embeddings in a fully unsupervised fashion, i.e. without parallel data. We present the first feasibility study, and investigate conformity of existing vision and language foundation models in the context of unsupervised, or "blind", matching. First, we formulate unsupervised matching as a quadratic assignment problem and introduce a novel heuristic that outperforms previous solvers. We also develop a technique to find optimal matching problems, for which a non-trivial match is very likely. Second, we conduct an extensive study deploying a range of vision and language models on four datasets. Our analysis reveals that for many problem instances, vision and language representations can be indeed matched without supervision. This finding opens up the exciting possibility of embedding semantic knowledge into other modalities virtually annotation-free. As a proof of concept, we showcase an unsupervised classifier, which achieves non-trivial classification accuracy without any image-text annotation.

comment: Accepted to CVPR 2025, Project page: https://dominik-schnaus.github.io/itsamatch/

☆ IMPACT: A Generic Semantic Loss for Multimodal Medical Image Registration

Valentin Boussot, Cédric Hémon, Jean-Claude Nunes, Jason Downling, Simon Rouzé, Caroline Lafond, Anaïs Barateau, Jean-Louis Dillenseger

Image registration is fundamental in medical imaging, enabling precise alignment of anatomical structures for diagnosis, treatment planning, image-guided treatment or longitudinal monitoring. This work introduces IMPACT (Image Metric with Pretrained model-Agnostic Comparison for Transmodality registration), a generic semantic similarity metric designed for seamless integration into diverse image registration frameworks (such as Elastix and Voxelmorph). It compares deep learning-based features extracted from medical images without requiring task-specific training, ensuring broad applicability across various modalities. By leveraging the features of the large-scale pretrained TotalSegmentator models and the ability to integrate Segment Anything Model (SAM) and other large-scale segmentation networks, this approach offers significant advantages. It provides robust, scalable, and efficient solutions for multimodal image registration. The IMPACT loss was evaluated on five challenging registration tasks involving thoracic CT/CBCT, and pelvic MR/CT datasets. Quantitative metrics, such as Target Registration Error and Dice Similarity Coefficient, demonstrated significant improvements in anatomical alignment compared to baseline methods. Qualitative analyses further confirmed the increased robustness of the proposed metric in the face of noise, artifacts, and modality variations. IMPACT's versatility and efficiency make it a valuable tool for advancing registration performance in clinical and research applications, addressing critical challenges in multimodal medical imaging.

comment: Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). This is a preprint version and has not been peer-reviewed

☆ PolypSegTrack: Unified Foundation Model for Colonoscopy Video Analysis

Anwesa Choudhuri, Zhongpai Gao, Meng Zheng, Benjamin Planche, Terrence Chen, Ziyan Wu

Early detection, accurate segmentation, classification and tracking of polyps during colonoscopy are critical for preventing colorectal cancer. Many existing deep-learning-based methods for analyzing colonoscopic videos either require task-specific fine-tuning, lack tracking capabilities, or rely on domain-specific pre-training. In this paper, we introduce \textit{PolypSegTrack}, a novel foundation model that jointly addresses polyp detection, segmentation, classification and unsupervised tracking in colonoscopic videos. Our approach leverages a novel conditional mask loss, enabling flexible training across datasets with either pixel-level segmentation masks or bounding box annotations, allowing us to bypass task-specific fine-tuning. Our unsupervised tracking module reliably associates polyp instances across frames using object queries, without relying on any heuristics. We leverage a robust vision foundation model backbone that is pre-trained unsupervisedly on natural images, thereby removing the need for domain-specific pre-training. Extensive experiments on multiple polyp benchmarks demonstrate that our method significantly outperforms existing state-of-the-art approaches in detection, segmentation, classification, and tracking.

☆ DANTE-AD: Dual-Vision Attention Network for Long-Term Audio Description

Adrienne Deganutti, Simon Hadfield, Andrew Gilbert

Audio Description is a narrated commentary designed to aid vision-impaired audiences in perceiving key visual elements in a video. While short-form video understanding has advanced rapidly, a solution for maintaining coherent long-term visual storytelling remains unresolved. Existing methods rely solely on frame-level embeddings, effectively describing object-based content but lacking contextual information across scenes. We introduce DANTE-AD, an enhanced video description model leveraging a dual-vision Transformer-based architecture to address this gap. DANTE-AD sequentially fuses both frame and scene level embeddings to improve long-term contextual understanding. We propose a novel, state-of-the-art method for sequential cross-attention to achieve contextual grounding for fine-grained audio description generation. Evaluated on a broad range of key scenes from well-known movie clips, DANTE-AD outperforms existing methods across traditional NLP metrics and LLM-based evaluations.

☆ 4D mmWave Radar in Adverse Environments for Autonomous Driving: A Survey

Xiangyuan Peng, Miao Tang, Huawei Sun, Lorenzo Servadei, Robert Wille

Autonomous driving systems require accurate and reliable perception. However, adverse environments, such as rain, snow, and fog, can significantly degrade the performance of LiDAR and cameras. In contrast, 4D millimeter-wave (mmWave) radar not only provides 3D sensing and additional velocity measurements but also maintains robustness in challenging conditions, making it increasingly valuable for autonomous driving. Recently, research on 4D mmWave radar under adverse environments has been growing, but a comprehensive survey is still lacking. To bridge this gap, this survey comprehensively reviews the current research on 4D mmWave radar under adverse environments. First, we present an overview of existing 4D mmWave radar datasets encompassing diverse weather and lighting scenarios. Next, we analyze methods and models according to different adverse conditions. Finally, the challenges faced in current studies and potential future directions are discussed for advancing 4D mmWave radar applications in harsh environments. To the best of our knowledge, this is the first survey specifically focusing on 4D mmWave radar in adverse environments for autonomous driving.

comment: 8 pages

☆ A Plasticity-Aware Method for Continual Self-Supervised Learning in Remote Sensing

Lars Möllenbrok, Behnood Rasti, Begüm Demir

Continual self-supervised learning (CSSL) methods have gained increasing attention in remote sensing (RS) due to their capability to learn new tasks sequentially from continuous streams of unlabeled data. Existing CSSL methods, while learning new tasks, focus on preventing catastrophic forgetting. To this end, most of them use regularization strategies to retain knowledge of previous tasks. This reduces the model's ability to adapt to the data of new tasks (i.e., learning plasticity), which can degrade performance. To address this problem, in this paper, we propose a novel CSSL method that aims to learn tasks sequentially, while achieving high learning plasticity. To this end, the proposed method uses a knowledge distillation strategy with an integrated decoupling mechanism. The decoupling is achieved by first dividing the feature dimensions into task-common and task-specific parts. Then, the task-common features are forced to be correlated to ensure memory stability while the task-specific features are forced to be de-correlated facilitating the learning of new features. Experimental results show the effectiveness of the proposed method compared to CaSSLe, which is a widely used CSSL framework, with improvements of up to 1.12% in average accuracy and 2.33% in intransigence in a task-incremental scenario, and 1.24% in average accuracy and 2.01% in intransigence in a class-incremental scenario.

comment: Accepted at IEEE International Geoscience and Remote Sensing Symposium 2025

☆ From Colors to Classes: Emergence of Concepts in Vision Transformers

Teresa Dorszewski, Lenka Tětková, Robert Jenssen, Lars Kai Hansen, Kristoffer Knutsen Wickstrøm

Vision Transformers (ViTs) are increasingly utilized in various computer vision tasks due to their powerful representation capabilities. However, it remains understudied how ViTs process information layer by layer. Numerous studies have shown that convolutional neural networks (CNNs) extract features of increasing complexity throughout their layers, which is crucial for tasks like domain adaptation and transfer learning. ViTs, lacking the same inductive biases as CNNs, can potentially learn global dependencies from the first layers due to their attention mechanisms. Given the increasing importance of ViTs in computer vision, there is a need to improve the layer-wise understanding of ViTs. In this work, we present a novel, layer-wise analysis of concepts encoded in state-of-the-art ViTs using neuron labeling. Our findings reveal that ViTs encode concepts with increasing complexity throughout the network. Early layers primarily encode basic features such as colors and textures, while later layers represent more specific classes, including objects and animals. As the complexity of encoded concepts increases, the number of concepts represented in each layer also rises, reflecting a more diverse and specific set of features. Additionally, different pretraining strategies influence the quantity and category of encoded concepts, with finetuning to specific downstream tasks generally reducing the number of encoded concepts and shifting the concepts to more relevant categories.

comment: Preprint. Accepted at The 3rd World Conference on eXplainable Artificial Intelligence

☆ COSMO: Combination of Selective Memorization for Low-cost Vision-and-Language Navigation

Siqi Zhang, Yanyuan Qiao, Qunbo Wang, Zike Yan, Qi Wu, Zhihua Wei, Jing Liu

Vision-and-Language Navigation (VLN) tasks have gained prominence within artificial intelligence research due to their potential application in fields like home assistants. Many contemporary VLN approaches, while based on transformer architectures, have increasingly incorporated additional components such as external knowledge bases or map information to enhance performance. These additions, while boosting performance, also lead to larger models and increased computational costs. In this paper, to achieve both high performance and low computational costs, we propose a novel architecture with the COmbination of Selective MemOrization (COSMO). Specifically, COSMO integrates state-space modules and transformer modules, and incorporates two VLN-customized selective state space modules: the Round Selective Scan (RSS) and the Cross-modal Selective State Space Module (CS3). RSS facilitates comprehensive inter-modal interactions within a single scan, while the CS3 module adapts the selective state space module into a dual-stream architecture, thereby enhancing the acquisition of cross-modal interactions. Experimental validations on three mainstream VLN benchmarks, REVERIE, R2R, and R2R-CE, not only demonstrate competitive navigation performance of our model but also show a significant reduction in computational costs.

☆ AMMSM: Adaptive Motion Magnification and Sparse Mamba for Micro-Expression Recognition ICME 2025

Xuxiong Liu, Tengteng Dong, Fei Wang, Weijie Feng, Xiao Sun

Micro-expressions are typically regarded as unconscious manifestations of a person's genuine emotions. However, their short duration and subtle signals pose significant challenges for downstream recognition. We propose a multi-task learning framework named the Adaptive Motion Magnification and Sparse Mamba (AMMSM) to address this. This framework aims to enhance the accurate capture of micro-expressions through self-supervised subtle motion magnification, while the sparse spatial selection Mamba architecture combines sparse activation with the advanced Visual Mamba model to model key motion regions and their valuable representations more effectively. Additionally, we employ evolutionary search to optimize the magnification factor and the sparsity ratios of spatial selection, followed by fine-tuning to improve performance further. Extensive experiments on two standard datasets demonstrate that the proposed AMMSM achieves state-of-the-art (SOTA) accuracy and robustness.

comment: Accepted by ICME 2025

☆ BBoxCut: A Targeted Data Augmentation Technique for Enhancing Wheat Head Detection Under Occlusions

Yasashwini Sai Gowri P, Karthik Seemakurthy, Andrews Agyemang Opoku, Sita Devi Bharatula

Wheat plays a critical role in global food security, making it one of the most extensively studied crops. Accurate identification and measurement of key characteristics of wheat heads are essential for breeders to select varieties for cross-breeding, with the goal of developing nutrient-dense, resilient, and sustainable cultivars. Traditionally, these measurements are performed manually, which is both time-consuming and inefficient. Advances in digital technologies have paved the way for automating this process. However, field conditions pose significant challenges, such as occlusions of leaves, overlapping wheat heads, varying lighting conditions, and motion blur. In this paper, we propose a novel data augmentation technique, BBoxCut, which uses random localized masking to simulate occlusions caused by leaves and neighboring wheat heads. We evaluated our approach using three state-of-the-art object detectors and observed mean average precision (mAP) gains of 2.76, 3.26, and 1.9 for Faster R-CNN, FCOS, and DETR, respectively. Our augmentation technique led to significant improvements both qualitatively and quantitatively. In particular, the improvements were particularly evident in scenarios involving occluded wheat heads, demonstrating the robustness of our method in challenging field conditions.

☆ HumanDreamer: Generating Controllable Human-Motion Videos via Decoupled Generation

Boyuan Wang, Xiaofeng Wang, Chaojun Ni, Guosheng Zhao, Zhiqin Yang, Zheng Zhu, Muyang Zhang, Yukun Zhou, Xinze Chen, Guan Huang, Lihong Liu, Xingang Wang

Human-motion video generation has been a challenging task, primarily due to the difficulty inherent in learning human body movements. While some approaches have attempted to drive human-centric video generation explicitly through pose control, these methods typically rely on poses derived from existing videos, thereby lacking flexibility. To address this, we propose HumanDreamer, a decoupled human video generation framework that first generates diverse poses from text prompts and then leverages these poses to generate human-motion videos. Specifically, we propose MotionVid, the largest dataset for human-motion pose generation. Based on the dataset, we present MotionDiT, which is trained to generate structured human-motion poses from text prompts. Besides, a novel LAMA loss is introduced, which together contribute to a significant improvement in FID by 62.4%, along with respective enhancements in R-precision for top1, top2, and top3 by 41.8%, 26.3%, and 18.3%, thereby advancing both the Text-to-Pose control accuracy and FID metrics. Our experiments across various Pose-to-Video baselines demonstrate that the poses generated by our method can produce diverse and high-quality human-motion videos. Furthermore, our model can facilitate other downstream tasks, such as pose sequence prediction and 2D-3D motion lifting.

comment: Project Page: https://humandreamer.github.io

☆ Crossmodal Knowledge Distillation with WordNet-Relaxed Text Embeddings for Robust Image Classification

Chenqi Guo, Mengshuo Rong, Qianli Feng, Rongfan Feng, Yinglong Ma

Crossmodal knowledge distillation (KD) aims to enhance a unimodal student using a multimodal teacher model. In particular, when the teacher's modalities include the student's, additional complementary information can be exploited to improve knowledge transfer. In supervised image classification, image datasets typically include class labels that represent high-level concepts, suggesting a natural avenue to incorporate textual cues for crossmodal KD. However, these labels rarely capture the deeper semantic structures in real-world visuals and can lead to label leakage if used directly as inputs, ultimately limiting KD performance. To address these issues, we propose a multi-teacher crossmodal KD framework that integrates CLIP image embeddings with learnable WordNet-relaxed text embeddings under a hierarchical loss. By avoiding direct use of exact class names and instead using semantically richer WordNet expansions, we mitigate label leakage and introduce more diverse textual cues. Experiments show that this strategy significantly boosts student performance, whereas noisy or overly precise text embeddings hinder distillation efficiency. Interpretability analyses confirm that WordNet-relaxed prompts encourage heavier reliance on visual features over textual shortcuts, while still effectively incorporating the newly introduced textual cues. Our method achieves state-of-the-art or second-best results on six public datasets, demonstrating its effectiveness in advancing crossmodal KD.

☆ Optimization of Layer Skipping and Frequency Scaling for Convolutional Neural Networks under Latency Constraint ECCV

Minh David Thao Chan, Ruoyu Zhao, Yukuan Jia, Ruiqing Mao, Sheng Zhou

The energy consumption of Convolutional Neural Networks (CNNs) is a critical factor in deploying deep learning models on resource-limited equipment such as mobile devices and autonomous vehicles. We propose an approach involving Proportional Layer Skipping (PLS) and Frequency Scaling (FS). Layer skipping reduces computational complexity by selectively bypassing network layers, whereas frequency scaling adjusts the frequency of the processor to optimize energy use under latency constraints. Experiments of PLS and FS on ResNet-152 with the CIFAR-10 dataset demonstrated significant reductions in computational demands and energy consumption with minimal accuracy loss. This study offers practical solutions for improving real-time processing in resource-limited settings and provides insights into balancing computational efficiency and model performance.

comment: 12 pages, 6 figures, Accepted in Proc. Eur. Conf. Comput. Vis. (ECCV) Workshops. Milan, Italy: Springer, September 2024

☆ Learning 3D-Gaussian Simulators from RGB Videos

Mikel Zhobro, Andreas René Geist, Georg Martius

Learning physics simulations from video data requires maintaining spatial and temporal consistency, a challenge often addressed with strong inductive biases or ground-truth 3D information -- limiting scalability and generalization. We introduce 3DGSim, a 3D physics simulator that learns object dynamics end-to-end from multi-view RGB videos. It encodes images into a 3D Gaussian particle representation, propagates dynamics via a transformer, and renders frames using 3D Gaussian splatting. By jointly training inverse rendering with a dynamics transformer using a temporal encoding and merging layer, 3DGSimembeds physical properties into point-wise latent vectors without enforcing explicit connectivity constraints. This enables the model to capture diverse physical behaviors, from rigid to elastic and cloth-like interactions, along with realistic lighting effects that also generalize to unseen multi-body interactions and novel scene edits.

☆ H2VU-Benchmark: A Comprehensive Benchmark for Hierarchical Holistic Video Understanding

Qi Wu, Quanlong Zheng, Yanhao Zhang, Junlin Xie, Jinguo Luo, Kuo Wang, Peng Liu, Qingsong Xie, Ru Zhen, Haonan Lu, Zhenyu Yang

With the rapid development of multimodal models, the demand for assessing video understanding capabilities has been steadily increasing. However, existing benchmarks for evaluating video understanding exhibit significant limitations in coverage, task diversity, and scene adaptability. These shortcomings hinder the accurate assessment of models' comprehensive video understanding capabilities. To tackle this challenge, we propose a hierarchical and holistic video understanding (H2VU) benchmark designed to evaluate both general video and online streaming video comprehension. This benchmark contributes three key features: Extended video duration: Spanning videos from brief 3-second clips to comprehensive 1.5-hour recordings, thereby bridging the temporal gaps found in current benchmarks. Comprehensive assessment tasks: Beyond traditional perceptual and reasoning tasks, we have introduced modules for countercommonsense comprehension and trajectory state tracking. These additions test the models' deep understanding capabilities beyond mere prior knowledge. Enriched video data: To keep pace with the rapid evolution of current AI agents, we have expanded first-person streaming video datasets. This expansion allows for the exploration of multimodal models' performance in understanding streaming videos from a first-person perspective. Extensive results from H2VU reveal that existing multimodal large language models (MLLMs) possess substantial potential for improvement in our newly proposed evaluation tasks. We expect that H2VU will facilitate advancements in video understanding research by offering a comprehensive and in-depth analysis of MLLMs.

☆ DenseFormer: Learning Dense Depth Map from Sparse Depth and Image via Conditional Diffusion Model

Ming Yuan, Sichao Wang, Chuang Zhang, Lei He, Qing Xu, Jianqiang Wang

The depth completion task is a critical problem in autonomous driving, involving the generation of dense depth maps from sparse depth maps and RGB images. Most existing methods employ a spatial propagation network to iteratively refine the depth map after obtaining an initial dense depth. In this paper, we propose DenseFormer, a novel method that integrates the diffusion model into the depth completion task. By incorporating the denoising mechanism of the diffusion model, DenseFormer generates the dense depth map by progressively refining an initial random depth distribution through multiple iterations. We propose a feature extraction module that leverages a feature pyramid structure, along with multi-layer deformable attention, to effectively extract and integrate features from sparse depth maps and RGB images, which serve as the guiding condition for the diffusion process. Additionally, this paper presents a depth refinement module that applies multi-step iterative refinement across various ranges to the dense depth results generated by the diffusion process. The module utilizes image features enriched with multi-scale information and sparse depth input to further enhance the accuracy of the predicted depth map. Extensive experiments on the KITTI outdoor scene dataset demonstrate that DenseFormer outperforms classical depth completion methods.

☆ SALT: A Flexible Semi-Automatic Labeling Tool for General LiDAR Point Clouds with Cross-Scene Adaptability and 4D Consistency

Yanbo Wang, Yongtao Chen, Chuan Cao, Tianchen Deng, Wentao Zhao, Jingchuan Wang, Weidong Chen

We propose a flexible Semi-Automatic Labeling Tool (SALT) for general LiDAR point clouds with cross-scene adaptability and 4D consistency. Unlike recent approaches that rely on camera distillation, SALT operates directly on raw LiDAR data, automatically generating pre-segmentation results. To achieve this, we propose a novel zero-shot learning paradigm, termed data alignment, which transforms LiDAR data into pseudo-images by aligning with the training distribution of vision foundation models. Additionally, we design a 4D-consistent prompting strategy and 4D non-maximum suppression module to enhance SAM2, ensuring high-quality, temporally consistent presegmentation. SALT surpasses the latest zero-shot methods by 18.4% PQ on SemanticKITTI and achieves nearly 40-50% of human annotator performance on our newly collected low-resolution LiDAR data and on combined data from three LiDAR types, significantly boosting annotation efficiency. We anticipate that SALT's open-sourcing will catalyze substantial expansion of current LiDAR datasets and lay the groundwork for the future development of LiDAR foundation models. Code is available at https://github.com/Cavendish518/SALT.

☆ Video-based Traffic Light Recognition by Rockchip RV1126 for Autonomous Driving

Miao Fan, Xuxu Kong, Shengtong Xu, Haoyi Xiong, Xiangzeng Liu

Real-time traffic light recognition is fundamental for autonomous driving safety and navigation in urban environments. While existing approaches rely on single-frame analysis from onboard cameras, they struggle with complex scenarios involving occlusions and adverse lighting conditions. We present \textit{ViTLR}, a novel video-based end-to-end neural network that processes multiple consecutive frames to achieve robust traffic light detection and state classification. The architecture leverages a transformer-like design with convolutional self-attention modules, which is optimized specifically for deployment on the Rockchip RV1126 embedded platform. Extensive evaluations on two real-world datasets demonstrate that \textit{ViTLR} achieves state-of-the-art performance while maintaining real-time processing capabilities (>25 FPS) on RV1126's NPU. The system shows superior robustness across temporal stability, varying target distances, and challenging environmental conditions compared to existing single-frame approaches. We have successfully integrated \textit{ViTLR} into an ego-lane traffic light recognition system using HD maps for autonomous driving applications. The complete implementation, including source code and datasets, is made publicly available to facilitate further research in this domain.

comment: Accepted by IEEE IV'25

☆ A Benchmark for Vision-Centric HD Mapping by V2I Systems

Miao Fan, Shanshan Yu, Shengtong Xu, Kun Jiang, Haoyi Xiong, Xiangzeng Liu

Autonomous driving faces safety challenges due to a lack of global perspective and the semantic information of vectorized high-definition (HD) maps. Information from roadside cameras can greatly expand the map perception range through vehicle-to-infrastructure (V2I) communications. However, there is still no dataset from the real world available for the study on map vectorization onboard under the scenario of vehicle-infrastructure cooperation. To prosper the research on online HD mapping for Vehicle-Infrastructure Cooperative Autonomous Driving (VICAD), we release a real-world dataset, which contains collaborative camera frames from both vehicles and roadside infrastructures, and provides human annotations of HD map elements. We also present an end-to-end neural framework (i.e., V2I-HD) leveraging vision-centric V2I systems to construct vectorized maps. To reduce computation costs and further deploy V2I-HD on autonomous vehicles, we introduce a directionally decoupled self-attention mechanism to V2I-HD. Extensive experiments show that V2I-HD has superior performance in real-time inference speed, as tested by our real-world dataset. Abundant qualitative results also demonstrate stable and robust map construction quality with low cost in complex and various driving scenes. As a benchmark, both source codes and the dataset have been released at OneDrive for the purpose of further study.

comment: Accepted by IEEE IV'25

☆ Local Information Matters: Inference Acceleration For Grounded Conversation Generation Models Through Adaptive Local-Aware Token Pruning

Bizhe Bai, Jianjian Cao, Yadan Luo, Tao Che

Grounded Conversation Generation (GCG) is an emerging vision-language task that requires models to generate natural language responses seamlessly intertwined with corresponding object segmentation masks. Recent models, such as GLaMM and OMG-LLaVA, achieve pixel-level grounding but incur significant computational costs due to processing a large number of visual tokens. Existing token pruning methods, like FastV and PyramidDrop, fail to preserve the local visual features critical for accurate grounding, leading to substantial performance drops in GCG tasks. To address this, we propose Adaptive Local-Aware Token Pruning (ALTP), a simple yet effective framework that accelerates GCG models by prioritizing local object information. ALTP introduces two key components: (1) Detail Density Capture (DDC), which uses superpixel segmentation to retain tokens in object-centric regions, preserving fine-grained details, and (2) Dynamic Density Formation (DDF), which dynamically allocates tokens based on information density, ensuring higher retention in semantically rich areas. Extensive experiments on the GranDf dataset demonstrate that ALTP significantly outperforms existing token pruning methods, such as FastV and PyramidDrop, on both GLaMM and OMG-LLaVA models. Notably, when applied to GLaMM, ALTP achieves a 90% reduction in visual tokens with a 4.9% improvement in AP50 and a 5.0% improvement in Recall compared to PyramidDrop. Similarly, on OMG-LLaVA, ALTP improves AP by 2.1% and mIOU by 3.0% at a 90% token reduction compared with PDrop.

comment: Work in progress

☆ A Multi-Stage Auto-Context Deep Learning Framework for Tissue and Nuclei Segmentation and Classification in H&E-Stained Histological Images of Advanced Melanoma

Nima Torbati, Anastasia Meshcheryakova, Diana Mechtcheriakova, Amirreza Mahbod

Melanoma is the most lethal form of skin cancer, with an increasing incidence rate worldwide. Analyzing histological images of melanoma by localizing and classifying tissues and cell nuclei is considered the gold standard method for diagnosis and treatment options for patients. While many computerized approaches have been proposed for automatic analysis, most perform tissue-based analysis and nuclei (cell)-based analysis as separate tasks, which might be suboptimal. In this work, using the PUMA challenge dataset, we proposed a novel multi-stage deep learning approach by combining tissue and nuclei information in a unified framework based on the auto-context concept to perform segmentation and classification in histological images of melanoma. Through pre-training and further post-processing, our approach achieved second and first place rankings in the PUMA challenge, with average micro Dice tissue score and summed nuclei F1-score of 73.40% for Track 1 and 63.48% for Track 2, respectively. Our implementation for training and testing is available at: https://github.com/NimaTorbati/PumaSubmit

comment: 15 pages

☆ AirCache: Activating Inter-modal Relevancy KV Cache Compression for Efficient Large Vision-Language Model Inference

Kai Huang, Hao Zou, Bochen Wang, Ye Xi, Zhen Xie, Hao Wang

Recent advancements in Large Visual Language Models (LVLMs) have gained significant attention due to their remarkable reasoning capabilities and proficiency in generalization. However, processing a large number of visual tokens and generating long-context outputs impose substantial computational overhead, leading to excessive demands for key-value (KV) cache. To address this critical bottleneck, we propose AirCache, a novel KV cache compression method aimed at accelerating LVLMs inference. This work systematically investigates the correlations between visual and textual tokens within the attention mechanisms of LVLMs. Our empirical analysis reveals considerable redundancy in cached visual tokens, wherein strategically eliminating these tokens preserves model performance while significantly accelerating context generation. Inspired by these findings, we introduce an elite observation window for assessing the importance of visual components in the KV cache, focusing on stable inter-modal relevancy modeling with enhanced multi-perspective consistency. Additionally, we develop an adaptive layer-wise budget allocation strategy that capitalizes on the strength and skewness of token importance distribution, showcasing superior efficiency compared to uniform allocation. Comprehensive evaluations across multiple LVLMs and benchmarks demonstrate that our method achieves comparable performance to the full cache while retaining only 10% of visual KV cache, thereby reducing decoding latency by 29% to 66% across various batch size and prompt length of inputs. Notably, as cache retention rates decrease, our method exhibits increasing performance advantages over existing approaches.

☆ JointTuner: Appearance-Motion Adaptive Joint Training for Customized Video Generation

Fangda Chen, Shanshan Zhao, Chuanfu Xu, Long Lan

Recent text-to-video advancements have enabled coherent video synthesis from prompts and expanded to fine-grained control over appearance and motion. However, existing methods either suffer from concept interference due to feature domain mismatch caused by naive decoupled optimizations or exhibit appearance contamination induced by spatial feature leakage resulting from the entanglement of motion and appearance in reference video reconstructions. In this paper, we propose JointTuner, a novel adaptive joint training framework, to alleviate these issues. Specifically, we develop Adaptive LoRA, which incorporates a context-aware gating mechanism, and integrate the gated LoRA components into the spatial and temporal Transformers within the diffusion model. These components enable simultaneous optimization of appearance and motion, eliminating concept interference. In addition, we introduce the Appearance-independent Temporal Loss, which decouples motion patterns from intrinsic appearance in reference video reconstructions through an appearance-agnostic noise prediction task. The key innovation lies in adding frame-wise offset noise to the ground-truth Gaussian noise, perturbing its distribution, thereby disrupting spatial attributes associated with frames while preserving temporal coherence. Furthermore, we construct a benchmark comprising 90 appearance-motion customized combinations and 10 multi-type automatic metrics across four dimensions, facilitating a more comprehensive evaluation for this customization task. Extensive experiments demonstrate the superior performance of our method compared to current advanced approaches.

comment: Project Page: https://fdchen24.github.io/JointTuner-Website

☆ AMB-FHE: Adaptive Multi-biometric Fusion with Fully Homomorphic Encryption

Florian Bayer, Christian Rathgeb

Biometric systems strive to balance security and usability. The use of multi-biometric systems combining multiple biometric modalities is usually recommended for high-security applications. However, the presentation of multiple biometric modalities can impair the user-friendliness of the overall system and might not be necessary in all cases. In this work, we present a simple but flexible approach to increase the privacy protection of homomorphically encrypted multi-biometric reference templates while enabling adaptation to security requirements at run-time: An adaptive multi-biometric fusion with fully homomorphic encryption (AMB-FHE). AMB-FHE is benchmarked against a bimodal biometric database consisting of the CASIA iris and MCYT fingerprint datasets using deep neural networks for feature extraction. Our contribution is easy to implement and increases the flexibility of biometric authentication while offering increased privacy protection through joint encryption of templates from multiple modalities.

☆ Spectral-Adaptive Modulation Networks for Visual Perception

Guhnoo Yun, Juhan Yoo, Kijung Kim, Jeongho Lee, Paul Hongsuck Seo, Dong Hwan Kim

Recent studies have shown that 2D convolution and self-attention exhibit distinct spectral behaviors, and optimizing their spectral properties can enhance vision model performance. However, theoretical analyses remain limited in explaining why 2D convolution is more effective in high-pass filtering than self-attention and why larger kernels favor shape bias, akin to self-attention. In this paper, we employ graph spectral analysis to theoretically simulate and compare the frequency responses of 2D convolution and self-attention within a unified framework. Our results corroborate previous empirical findings and reveal that node connectivity, modulated by window size, is a key factor in shaping spectral functions. Leveraging this insight, we introduce a \textit{spectral-adaptive modulation} (SPAM) mixer, which processes visual features in a spectral-adaptive manner using multi-scale convolutional kernels and a spectral re-scaling mechanism to refine spectral components. Based on SPAM, we develop SPANetV2 as a novel vision backbone. Extensive experiments demonstrate that SPANetV2 outperforms state-of-the-art models across multiple vision tasks, including ImageNet-1K classification, COCO object detection, and ADE20K semantic segmentation.

☆ Exploring Reliable PPG Authentication on Smartwatches in Daily Scenarios

Jiankai Tang, Jiacheng Liu, Renling Tong, Kai Zhu, Zhe Li, Xin Yi, Junliang Xing, Yuanchun Shi, Yuntao Wang

Photoplethysmography (PPG) Sensors, widely deployed in smartwatches, offer a simple and non-invasive authentication approach for daily use. However, PPG authentication faces reliability issues due to motion artifacts from physical activity and physiological variability over time. To address these challenges, we propose MTL-RAPID, an efficient and reliable PPG authentication model, that employs a multitask joint training strategy, simultaneously assessing signal quality and verifying user identity. The joint optimization of these two tasks in MTL-RAPID results in a structure that outperforms models trained on individual tasks separately, achieving stronger performance with fewer parameters. In our comprehensive user studies regarding motion artifacts (N = 30), time variations (N = 32), and user preferences (N = 16), MTL-RAPID achieves a best AUC of 99.2\% and an EER of 3.5\%, outperforming existing baselines. We opensource our PPG authentication dataset along with the MTL-RAPID model to facilitate future research on GitHub.

☆ CoMatch: Dynamic Covisibility-Aware Transformer for Bilateral Subpixel-Level Semi-Dense Image Matching

Zizhuo Li, Yifan Lu, Linfeng Tang, Shihua Zhang, Jiayi Ma

This prospective study proposes CoMatch, a novel semi-dense image matcher with dynamic covisibility awareness and bilateral subpixel accuracy. Firstly, observing that modeling context interaction over the entire coarse feature map elicits highly redundant computation due to the neighboring representation similarity of tokens, a covisibility-guided token condenser is introduced to adaptively aggregate tokens in light of their covisibility scores that are dynamically estimated, thereby ensuring computational efficiency while improving the representational capacity of aggregated tokens simultaneously. Secondly, considering that feature interaction with massive non-covisible areas is distracting, which may degrade feature distinctiveness, a covisibility-assisted attention mechanism is deployed to selectively suppress irrelevant message broadcast from non-covisible reduced tokens, resulting in robust and compact attention to relevant rather than all ones. Thirdly, we find that at the fine-level stage, current methods adjust only the target view's keypoints to subpixel level, while those in the source view remain restricted at the coarse level and thus not informative enough, detrimental to keypoint location-sensitive usages. A simple yet potent fine correlation module is developed to refine the matching candidates in both source and target views to subpixel level, attaining attractive performance improvement. Thorough experimentation across an array of public benchmarks affirms CoMatch's promising accuracy, efficiency, and generalizability.

☆ FineCausal: A Causal-Based Framework for Interpretable Fine-Grained Action Quality Assessment

Ruisheng Han, Kanglei Zhou, Amir Atapour-Abarghouei, Xiaohui Liang, Hubert P. H. Shum

Action quality assessment (AQA) is critical for evaluating athletic performance, informing training strategies, and ensuring safety in competitive sports. However, existing deep learning approaches often operate as black boxes and are vulnerable to spurious correlations, limiting both their reliability and interpretability. In this paper, we introduce FineCausal, a novel causal-based framework that achieves state-of-the-art performance on the FineDiving-HM dataset. Our approach leverages a Graph Attention Network-based causal intervention module to disentangle human-centric foreground cues from background confounders, and incorporates a temporal causal attention module to capture fine-grained temporal dependencies across action stages. This dual-module strategy enables FineCausal to generate detailed spatio-temporal representations that not only achieve state-of-the-art scoring performance but also provide transparent, interpretable feedback on which features drive the assessment. Despite its strong performance, FineCausal requires extensive expert knowledge to define causal structures and depends on high-quality annotations, challenges that we discuss and address as future research directions. Code is available at https://github.com/Harrison21/FineCausal.

☆ HumanAesExpert: Advancing a Multi-Modality Foundation Model for Human Image Aesthetic Assessment

Zhichao Liao, Xiaokun Liu, Wenyu Qin, Qingyu Li, Qiulin Wang, Pengfei Wan, Di Zhang, Long Zeng, Pingfa Feng

Image Aesthetic Assessment (IAA) is a long-standing and challenging research task. However, its subset, Human Image Aesthetic Assessment (HIAA), has been scarcely explored, even though HIAA is widely used in social media, AI workflows, and related domains. To bridge this research gap, our work pioneers a holistic implementation framework tailored for HIAA. Specifically, we introduce HumanBeauty, the first dataset purpose-built for HIAA, which comprises 108k high-quality human images with manual annotations. To achieve comprehensive and fine-grained HIAA, 50K human images are manually collected through a rigorous curation process and annotated leveraging our trailblazing 12-dimensional aesthetic standard, while the remaining 58K with overall aesthetic labels are systematically filtered from public datasets. Based on the HumanBeauty database, we propose HumanAesExpert, a powerful Vision Language Model for aesthetic evaluation of human images. We innovatively design an Expert head to incorporate human knowledge of aesthetic sub-dimensions while jointly utilizing the Language Modeling (LM) and Regression head. This approach empowers our model to achieve superior proficiency in both overall and fine-grained HIAA. Furthermore, we introduce a MetaVoter, which aggregates scores from all three heads, to effectively balance the capabilities of each head, thereby realizing improved assessment precision. Extensive experiments demonstrate that our HumanAesExpert models deliver significantly better performance in HIAA than other state-of-the-art models. Our datasets, models, and codes are publicly released to advance the HIAA community. Project webpage: https://humanaesexpert.github.io/HumanAesExpert/

☆ Boosting MLLM Reasoning with Text-Debiased Hint-GRPO

Qihan Huang, Long Chan, Jinlong Liu, Wanggui He, Hao Jiang, Mingli Song, Jingyuan Chen, Chang Yao, Jie Song

MLLM reasoning has drawn widespread research for its excellent problem-solving capability. Current reasoning methods fall into two types: PRM, which supervises the intermediate reasoning steps, and ORM, which supervises the final results. Recently, DeepSeek-R1 has challenged the traditional view that PRM outperforms ORM, which demonstrates strong generalization performance using an ORM method (i.e., GRPO). However, current MLLM's GRPO algorithms still struggle to handle challenging and complex multimodal reasoning tasks (e.g., mathematical reasoning). In this work, we reveal two problems that impede the performance of GRPO on the MLLM: Low data utilization and Text-bias. Low data utilization refers to that GRPO cannot acquire positive rewards to update the MLLM on difficult samples, and text-bias is a phenomenon that the MLLM bypasses image condition and solely relies on text condition for generation after GRPO training. To tackle these problems, this work proposes Hint-GRPO that improves data utilization by adaptively providing hints for samples of varying difficulty, and text-bias calibration that mitigates text-bias by calibrating the token prediction logits with image condition in test-time. Experiment results on three base MLLMs across eleven datasets demonstrate that our proposed methods advance the reasoning capability of original MLLM by a large margin, exhibiting superior performance to existing MLLM reasoning methods. Our code is available at https://github.com/hqhQAQ/Hint-GRPO.

☆ An Explainable Neural Radiomic Sequence Model with Spatiotemporal Continuity for Quantifying 4DCT-based Pulmonary Ventilation

Rihui Zhang, Haiming Zhu, Jingtong Zhao, Lei Zhang, Fang-Fang Yin, Chunhao Wang, Zhenyu Yang

Accurate evaluation of regional lung ventilation is essential for the management and treatment of lung cancer patients, supporting assessments of pulmonary function, optimization of therapeutic strategies, and monitoring of treatment response. Currently, ventilation scintigraphy using nuclear medicine techniques is widely employed in clinical practice; however, it is often time-consuming, costly, and entails additional radiation exposure. In this study, we propose an explainable neural radiomic sequence model to identify regions of compromised pulmonary ventilation based on four-dimensional computed tomography (4DCT). A cohort of 45 lung cancer patients from the VAMPIRE dataset was analyzed. For each patient, lung volumes were segmented from 4DCT, and voxel-wise radiomic features (56-dimensional) were extracted across the respiratory cycle to capture local intensity and texture dynamics, forming temporal radiomic sequences. Ground truth ventilation defects were delineated voxel-wise using Galligas-PET and DTPA-SPECT. To identify compromised regions, we developed a temporal saliency-enhanced explainable long short-term memory (LSTM) network trained on the radiomic sequences. Temporal saliency maps were generated to highlight key features contributing to the model's predictions. The proposed model demonstrated robust performance, achieving average (range) Dice similarity coefficients of 0.78 (0.74-0.79) for 25 PET cases and 0.78 (0.74-0.82) for 20 SPECT cases. The temporal saliency map explained three key radiomic sequences in ventilation quantification: during lung exhalation, compromised pulmonary function region typically exhibits (1) an increasing trend of intensity and (2) a decreasing trend of homogeneity, in contrast to healthy lung tissue.

comment: 43 pages, 13 figures

☆ Training-Free Text-Guided Image Editing with Visual Autoregressive Model

Yufei Wang, Lanqing Guo, Zhihao Li, Jiaxing Huang, Pichao Wang, Bihan Wen, Jian Wang

Text-guided image editing is an essential task that enables users to modify images through natural language descriptions. Recent advances in diffusion models and rectified flows have significantly improved editing quality, primarily relying on inversion techniques to extract structured noise from input images. However, inaccuracies in inversion can propagate errors, leading to unintended modifications and compromising fidelity. Moreover, even with perfect inversion, the entanglement between textual prompts and image features often results in global changes when only local edits are intended. To address these challenges, we propose a novel text-guided image editing framework based on VAR (Visual AutoRegressive modeling), which eliminates the need for explicit inversion while ensuring precise and controlled modifications. Our method introduces a caching mechanism that stores token indices and probability distributions from the original image, capturing the relationship between the source prompt and the image. Using this cache, we design an adaptive fine-grained masking strategy that dynamically identifies and constrains modifications to relevant regions, preventing unintended changes. A token reassembling approach further refines the editing process, enhancing diversity, fidelity, and control. Our framework operates in a training-free manner and achieves high-fidelity editing with faster inference speeds, processing a 1K resolution image in as fast as 1.2 seconds. Extensive experiments demonstrate that our method achieves performance comparable to, or even surpassing, existing diffusion- and rectified flow-based approaches in both quantitative metrics and visual quality. The code will be released.

☆ DiffScale: Continuous Downscaling and Bias Correction of Subseasonal Wind Speed Forecasts using Diffusion Models

Maximilian Springenberg, Noelia Otero, Yuxin Xue, Jackie Ma

Renewable resources are strongly dependent on local and large-scale weather situations. Skillful subseasonal to seasonal (S2S) forecasts -- beyond two weeks and up to two months -- can offer significant socioeconomic advantages to the energy sector. This study aims to enhance wind speed predictions using a diffusion model with classifier-free guidance to downscale S2S forecasts of surface wind speed. We propose DiffScale, a diffusion model that super-resolves spatial information for continuous downscaling factors and lead times. Leveraging weather priors as guidance for the generative process of diffusion models, we adopt the perspective of conditional probabilities on sampling super-resolved S2S forecasts. We aim to directly estimate the density associated with the target S2S forecasts at different spatial resolutions and lead times without auto-regression or sequence prediction, resulting in an efficient and flexible model. Synthetic experiments were designed to super-resolve wind speed S2S forecasts from the European Center for Medium-Range Weather Forecast (ECMWF) from a coarse resolution to a finer resolution of ERA5 reanalysis data, which serves as a high-resolution target. The innovative aspect of DiffScale lies in its flexibility to downscale arbitrary scaling factors, enabling it to generalize across various grid resolutions and lead times -without retraining the model- while correcting model errors, making it a versatile tool for improving S2S wind speed forecasts. We achieve a significant improvement in prediction quality, outperforming baselines up to week 3.

comment: 28 pages, 18 figures, preprint under review

☆ MuseFace: Text-driven Face Editing via Diffusion-based Mask Generation Approach

Xin Zhang, Siting Huang, Xiangyang Luo, Yifan Xie, Weijiang Yu, Heng Chang, Fei Ma, Fei Yu

Face editing modifies the appearance of face, which plays a key role in customization and enhancement of personal images. Although much work have achieved remarkable success in text-driven face editing, they still face significant challenges as none of them simultaneously fulfill the characteristics of diversity, controllability and flexibility. To address this challenge, we propose MuseFace, a text-driven face editing framework, which relies solely on text prompt to enable face editing. Specifically, MuseFace integrates a Text-to-Mask diffusion model and a semantic-aware face editing model, capable of directly generating fine-grained semantic masks from text and performing face editing. The Text-to-Mask diffusion model provides \textit{diversity} and \textit{flexibility} to the framework, while the semantic-aware face editing model ensures \textit{controllability} of the framework. Our framework can create fine-grained semantic masks, making precise face editing possible, and significantly enhancing the controllability and flexibility of face editing models. Extensive experiments demonstrate that MuseFace achieves superior high-fidelity performance.

comment: 6 pages, 5 figures,IEEE International Conference on Multimedia & Expo 2025

☆ GLane3D : Detecting Lanes with Graph of 3D Keypoints CVPR 2025

Halil İbrahim Öztürk, Muhammet Esat Kalfaoğlu, Ozsel Kilinc

Accurate and efficient lane detection in 3D space is essential for autonomous driving systems, where robust generalization is the foremost requirement for 3D lane detection algorithms. Considering the extensive variation in lane structures worldwide, achieving high generalization capacity is particularly challenging, as algorithms must accurately identify a wide variety of lane patterns worldwide. Traditional top-down approaches rely heavily on learning lane characteristics from training datasets, often struggling with lanes exhibiting previously unseen attributes. To address this generalization limitation, we propose a method that detects keypoints of lanes and subsequently predicts sequential connections between them to construct complete 3D lanes. Each key point is essential for maintaining lane continuity, and we predict multiple proposals per keypoint by allowing adjacent grids to predict the same keypoint using an offset mechanism. PointNMS is employed to eliminate overlapping proposal keypoints, reducing redundancy in the estimated BEV graph and minimizing computational overhead from connection estimations. Our model surpasses previous state-of-the-art methods on both the Apollo and OpenLane datasets, demonstrating superior F1 scores and a strong generalization capacity when models trained on OpenLane are evaluated on the Apollo dataset, compared to prior approaches.

comment: Accepted to CVPR 2025

☆ ExScene: Free-View 3D Scene Reconstruction with Gaussian Splatting from a Single Image ICME 2025

Tianyi Gong, Boyan Li, Yifei Zhong, Fangxin Wang

The increasing demand for augmented and virtual reality applications has highlighted the importance of crafting immersive 3D scenes from a simple single-view image. However, due to the partial priors provided by single-view input, existing methods are often limited to reconstruct low-consistency 3D scenes with narrow fields of view from single-view input. These limitations make them less capable of generalizing to reconstruct immersive scenes. To address this problem, we propose ExScene, a two-stage pipeline to reconstruct an immersive 3D scene from any given single-view image. ExScene designs a novel multimodal diffusion model to generate a high-fidelity and globally consistent panoramic image. We then develop a panoramic depth estimation approach to calculate geometric information from panorama, and we combine geometric information with high-fidelity panoramic image to train an initial 3D Gaussian Splatting (3DGS) model. Following this, we introduce a GS refinement technique with 2D stable video diffusion priors. We add camera trajectory consistency and color-geometric priors into the denoising process of diffusion to improve color and spatial consistency across image sequences. These refined sequences are then used to fine-tune the initial 3DGS model, leading to better reconstruction quality. Experimental results demonstrate that our ExScene achieves consistent and immersive scene reconstruction using only single-view input, significantly surpassing state-of-the-art baselines.

comment: ICME 2025

☆ ZeroMimic: Distilling Robotic Manipulation Skills from Web Videos ICRA 2025

Junyao Shi, Zhuolun Zhao, Tianyou Wang, Ian Pedroza, Amy Luo, Jie Wang, Jason Ma, Dinesh Jayaraman

Many recent advances in robotic manipulation have come through imitation learning, yet these rely largely on mimicking a particularly hard-to-acquire form of demonstrations: those collected on the same robot in the same room with the same objects as the trained policy must handle at test time. In contrast, large pre-recorded human video datasets demonstrating manipulation skills in-the-wild already exist, which contain valuable information for robots. Is it possible to distill a repository of useful robotic skill policies out of such data without any additional requirements on robot-specific demonstrations or exploration? We present the first such system ZeroMimic, that generates immediately deployable image goal-conditioned skill policies for several common categories of manipulation tasks (opening, closing, pouring, pick&place, cutting, and stirring) each capable of acting upon diverse objects and across diverse unseen task setups. ZeroMimic is carefully designed to exploit recent advances in semantic and geometric visual understanding of human videos, together with modern grasp affordance detectors and imitation policy classes. After training ZeroMimic on the popular EpicKitchens dataset of ego-centric human videos, we evaluate its out-of-the-box performance in varied real-world and simulated kitchen settings with two different robot embodiments, demonstrating its impressive abilities to handle these varied tasks. To enable plug-and-play reuse of ZeroMimic policies on other task setups and robots, we release software and policy checkpoints of our skill policies.

comment: ICRA 2025. Project website: https://zeromimic.github.io/

☆ Learned Image Compression and Restoration for Digital Pathology

SeonYeong Lee, EonSeung Seong, DongEon Lee, SiYeoul Lee, Yubin Cho, Chunsu Park, Seonho Kim, MinKyoung Seo, YoungSin Ko, MinWoo Kim

Digital pathology images play a crucial role in medical diagnostics, but their ultra-high resolution and large file sizes pose significant challenges for storage, transmission, and real-time visualization. To address these issues, we propose CLERIC, a novel deep learning-based image compression framework designed specifically for whole slide images (WSIs). CLERIC integrates a learnable lifting scheme and advanced convolutional techniques to enhance compression efficiency while preserving critical pathological details. Our framework employs a lifting-scheme transform in the analysis stage to decompose images into low- and high-frequency components, enabling more structured latent representations. These components are processed through parallel encoders incorporating Deformable Residual Blocks (DRB) and Recurrent Residual Blocks (R2B) to improve feature extraction and spatial adaptability. The synthesis stage applies an inverse lifting transform for effective image reconstruction, ensuring high-fidelity restoration of fine-grained tissue structures. We evaluate CLERIC on a digital pathology image dataset and compare its performance against state-of-the-art learned image compression (LIC) models. Experimental results demonstrate that CLERIC achieves superior rate-distortion (RD) performance, significantly reducing storage requirements while maintaining high diagnostic image quality. Our study highlights the potential of deep learning-based compression in digital pathology, facilitating efficient data management and long-term storage while ensuring seamless integration into clinical workflows and AI-assisted diagnostic systems. Code and models are available at: https://github.com/pnu-amilab/CLERIC.

☆ FlexiMo: A Flexible Remote Sensing Foundation Model

Xuyang Li, Chenyu Li, Pedram Ghamisi, Danfeng Hong

The rapid expansion of multi-source satellite imagery drives innovation in Earth observation, opening unprecedented opportunities for Remote Sensing Foundation Models to harness diverse data. However, many existing models remain constrained by fixed spatial resolutions and patch sizes, limiting their ability to fully exploit the heterogeneous spatial characteristics inherent in satellite imagery. To address these challenges, we propose FlexiMo, a flexible remote sensing foundation model that endows the pre-trained model with the flexibility to adapt to arbitrary spatial resolutions. Central to FlexiMo is a spatial resolution-aware module that employs a parameter-free alignment embedding mechanism to dynamically recalibrate patch embeddings based on the input image's resolution and dimensions. This design not only preserves critical token characteristics and ensures multi-scale feature fidelity but also enables efficient feature extraction without requiring modifications to the underlying network architecture. In addition, FlexiMo incorporates a lightweight channel adaptation module that leverages prior spectral information from sensors. This mechanism allows the model to process images with varying numbers of channels while maintaining the data's intrinsic physical properties. Extensive experiments on diverse multimodal, multi-resolution, and multi-scale datasets demonstrate that FlexiMo significantly enhances model generalization and robustness. In particular, our method achieves outstanding performance across a range of downstream tasks, including scene classification, land cover classification, urban building segmentation, and cloud detection. By enabling parameter-efficient and physically consistent adaptation, FlexiMo paves the way for more adaptable and effective foundation models in real-world remote sensing applications.

☆ Conformal uncertainty quantification to evaluate predictive fairness of foundation AI model for skin lesion classes across patient demographics

Swarnava Bhattacharyya, Umapada Pal, Tapabrata Chakraborti

Deep learning based diagnostic AI systems based on medical images are starting to provide similar performance as human experts. However these data hungry complex systems are inherently black boxes and therefore slow to be adopted for high risk applications like healthcare. This problem of lack of transparency is exacerbated in the case of recent large foundation models, which are trained in a self supervised manner on millions of data points to provide robust generalisation across a range of downstream tasks, but the embeddings generated from them happen through a process that is not interpretable, and hence not easily trustable for clinical applications. To address this timely issue, we deploy conformal analysis to quantify the predictive uncertainty of a vision transformer (ViT) based foundation model across patient demographics with respect to sex, age and ethnicity for the tasks of skin lesion classification using several public benchmark datasets. The significant advantage of this method is that conformal analysis is method independent and it not only provides a coverage guarantee at population level but also provides an uncertainty score for each individual. We used a model-agnostic dynamic F1-score-based sampling during model training, which helped to stabilize the class imbalance and we investigate the effects on uncertainty quantification (UQ) with or without this bias mitigation step. Thus we show how this can be used as a fairness metric to evaluate the robustness of the feature embeddings of the foundation model (Google DermFoundation) and thus advance the trustworthiness and fairness of clinical AI.

☆ Bridge the Gap Between Visual and Linguistic Comprehension for Generalized Zero-shot Semantic Segmentation

Xiaoqing Guo, Wuyang Li, Yixuan Yuan

Generalized zero-shot semantic segmentation (GZS3) aims to achieve the human-level capability of segmenting not only seen classes but also novel class regions unseen in the training data through introducing the bridge of semantic representations, e.g., word vector. While effective, the way of utilizing one semantic representation to associate the corresponding class and to enable the knowledge transfer from seen to unseen classes is insufficient as well as incompatible with human cognition. Inspired by the observation that humans often use some `part' and `state' information to comprehend the seen objects and imagine unseen classes, we decouple each class into detailed descriptions, including object parts and states. Based on the decoupling formulation, we propose a Decoupled Vision-Language Matching (DeVLMatch) framework, composed of spatial-part (SPMatch) and channel-state (CSMatch) matching modules, for GZS3. In SPMatch, we comprehend objects with spatial part information from both visual and linguistic perspectives and perform graph matching to bridge the gap. In CSMatch, states of objects from the linguistic perspective are matched to compatible channel information from the visual perspective. By decoupling and matching objects across visual and linguistic comprehension, we can explicitly introspect the relationship between seen and unseen classes in fine-grained object part and state levels, thereby facilitating the knowledge transfer from seen to unseen classes in visual space. The proposed DeVLMatch framework surpasses the previous GZS3 methods on standard benchmarks, including PASCAL VOC, COCO-Stuff, and CATARACTS, demonstrating its effectiveness.

☆ On-device Sora: Enabling Training-Free Diffusion-based Text-to-Video Generation for Mobile Devices

Bosung Kim, Kyuhwan Lee, Isu Jeong, Jungmin Cheon, Yeojin Lee, Seulki Lee

We present On-device Sora, the first model training-free solution for diffusion-based on-device text-to-video generation that operates efficiently on smartphone-grade devices. To address the challenges of diffusion-based text-to-video generation on computation- and memory-limited mobile devices, the proposed On-device Sora applies three novel techniques to pre-trained video generative models. First, Linear Proportional Leap (LPL) reduces the excessive denoising steps required in video diffusion through an efficient leap-based approach. Second, Temporal Dimension Token Merging (TDTM) minimizes intensive token-processing computation in attention layers by merging consecutive tokens along the temporal dimension. Third, Concurrent Inference with Dynamic Loading (CI-DL) dynamically partitions large models into smaller blocks and loads them into memory for concurrent model inference, effectively addressing the challenges of limited device memory. We implement On-device Sora on the iPhone 15 Pro, and the experimental evaluations show that it is capable of generating high-quality videos on the device, comparable to those produced by high-end GPUs. These results show that On-device Sora enables efficient and high-quality video generation on resource-constrained mobile devices. We envision the proposed On-device Sora as a significant first step toward democratizing state-of-the-art generative technologies, enabling video generation on commodity mobile and embedded devices without resource-intensive re-training for model optimization (compression). The code implementation is available at a GitHub repository(https://github.com/eai-lab/On-device-Sora).

☆ Pan-LUT: Efficient Pan-sharpening via Learnable Look-Up Tables

Zhongnan Cai, Yingying Wang, Yunlong Lin, Hui Zheng, Ge Meng, Zixu Lin, Jiaxin Xie, Junbin Lu, Yue Huang, Xinghao Ding

Recently, deep learning-based pan-sharpening algorithms have achieved notable advancements over traditional methods. However, many deep learning-based approaches incur substantial computational overhead during inference, especially with high-resolution images. This excessive computational demand limits the applicability of these methods in real-world scenarios, particularly in the absence of dedicated computing devices such as GPUs and TPUs. To address these challenges, we propose Pan-LUT, a novel learnable look-up table (LUT) framework for pan-sharpening that strikes a balance between performance and computational efficiency for high-resolution remote sensing images. To finely control the spectral transformation, we devise the PAN-guided look-up table (PGLUT) for channel-wise spectral mapping. To effectively capture fine-grained spatial details and adaptively learn local contexts, we introduce the spatial details look-up table (SDLUT) and adaptive aggregation look-up table (AALUT). Our proposed method contains fewer than 300K parameters and processes a 8K resolution image in under 1 ms using a single NVIDIA GeForce RTX 2080 Ti GPU, demonstrating significantly faster performance compared to other methods. Experiments reveal that Pan-LUT efficiently processes large remote sensing images in a lightweight manner, bridging the gap to real-world applications. Furthermore, our model surpasses SOTA methods in full-resolution scenes under real-world conditions, highlighting its effectiveness and efficiency.

comment: 12 pages, 6 figures

☆ MGD-SAM2: Multi-view Guided Detail-enhanced Segment Anything Model 2 for High-Resolution Class-agnostic Segmentation

Haoran Shen, Peixian Zhuang, Jiahao Kou, Yuxin Zeng, Haoying Xu, Jiangyun Li

Segment Anything Models (SAMs), as vision foundation models, have demonstrated remarkable performance across various image analysis tasks. Despite their strong generalization capabilities, SAMs encounter challenges in fine-grained detail segmentation for high-resolution class-independent segmentation (HRCS), due to the limitations in the direct processing of high-resolution inputs and low-resolution mask predictions, and the reliance on accurate manual prompts. To address these limitations, we propose MGD-SAM2 which integrates SAM2 with multi-view feature interaction between a global image and local patches to achieve precise segmentation. MGD-SAM2 incorporates the pre-trained SAM2 with four novel modules: the Multi-view Perception Adapter (MPAdapter), the Multi-view Complementary Enhancement Module (MCEM), the Hierarchical Multi-view Interaction Module (HMIM), and the Detail Refinement Module (DRM). Specifically, we first introduce MPAdapter to adapt the SAM2 encoder for enhanced extraction of local details and global semantics in HRCS images. Then, MCEM and HMIM are proposed to further exploit local texture and global context by aggregating multi-view features within and across multi-scales. Finally, DRM is designed to generate gradually restored high-resolution mask predictions, compensating for the loss of fine-grained details resulting from directly upsampling the low-resolution prediction maps. Experimental results demonstrate the superior performance and strong generalization of our model on multiple high-resolution and normal-resolution datasets. Code will be available at https://github.com/sevenshr/MGD-SAM2.

☆ Evaluation of (Un-)Supervised Machine Learning Methods for GNSS Interference Classification with Real-World Data Discrepancies

Lucas Heublein, Nisha L. Raichur, Tobias Feigl, Tobias Brieger, Fin Heuer, Lennart Asbach, Alexander Rügamer, Felix Ott

The accuracy and reliability of vehicle localization on roads are crucial for applications such as self-driving cars, toll systems, and digital tachographs. To achieve accurate positioning, vehicles typically use global navigation satellite system (GNSS) receivers to validate their absolute positions. However, GNSS-based positioning can be compromised by interference signals, necessitating the identification, classification, determination of purpose, and localization of such interference to mitigate or eliminate it. Recent approaches based on machine learning (ML) have shown superior performance in monitoring interference. However, their feasibility in real-world applications and environments has yet to be assessed. Effective implementation of ML techniques requires training datasets that incorporate realistic interference signals, including real-world noise and potential multipath effects that may occur between transmitter, receiver, and satellite in the operational area. Additionally, these datasets require reference labels. Creating such datasets is often challenging due to legal restrictions, as causing interference to GNSS sources is strictly prohibited. Consequently, the performance of ML-based methods in practical applications remains unclear. To address this gap, we describe a series of large-scale measurement campaigns conducted in real-world settings at two highway locations in Germany and the Seetal Alps in Austria, and in large-scale controlled indoor environments. We evaluate the latest supervised ML-based methods to report on their performance in real-world settings and present the applicability of pseudo-labeling for unsupervised learning. We demonstrate the challenges of combining datasets due to data discrepancies and evaluate outlier detection, domain adaptation, and data augmentation techniques to present the models' capabilities to adapt to changes in the datasets.

comment: 34 pages, 25 figures

☆ XLRS-Bench: Could Your Multimodal LLMs Understand Extremely Large Ultra-High-Resolution Remote Sensing Imagery? CVPR2025

Fengxiang Wang, Hongzhen Wang, Mingshuo Chen, Di Wang, Yulin Wang, Zonghao Guo, Qiang Ma, Long Lan, Wenjing Yang, Jing Zhang, Zhiyuan Liu, Maosong Sun

The astonishing breakthrough of multimodal large language models (MLLMs) has necessitated new benchmarks to quantitatively assess their capabilities, reveal their limitations, and indicate future research directions. However, this is challenging in the context of remote sensing (RS), since the imagery features ultra-high resolution that incorporates extremely complex semantic relationships. Existing benchmarks usually adopt notably smaller image sizes than real-world RS scenarios, suffer from limited annotation quality, and consider insufficient dimensions of evaluation. To address these issues, we present XLRS-Bench: a comprehensive benchmark for evaluating the perception and reasoning capabilities of MLLMs in ultra-high-resolution RS scenarios. XLRS-Bench boasts the largest average image size (8500$\times$8500) observed thus far, with all evaluation samples meticulously annotated manually, assisted by a novel semi-automatic captioner on ultra-high-resolution RS images. On top of the XLRS-Bench, 16 sub-tasks are defined to evaluate MLLMs' 10 kinds of perceptual capabilities and 6 kinds of reasoning capabilities, with a primary emphasis on advanced cognitive processes that facilitate real-world decision-making and the capture of spatiotemporal changes. The results of both general and RS-focused MLLMs on XLRS-Bench indicate that further efforts are needed for real-world RS applications. We have open-sourced XLRS-Bench to support further research in developing more powerful MLLMs for remote sensing.

comment: It has been accepted by CVPR2025

☆ Texture or Semantics? Vision-Language Models Get Lost in Font Recognition

Zhecheng Li, Guoxian Song, Yujun Cai, Zhen Xiong, Junsong Yuan, Yiwei Wang

Modern Vision-Language Models (VLMs) exhibit remarkable visual and linguistic capabilities, achieving impressive performance in various tasks such as image recognition and object localization. However, their effectiveness in fine-grained tasks remains an open question. In everyday scenarios, individuals encountering design materials, such as magazines, typography tutorials, research papers, or branding content, may wish to identify aesthetically pleasing fonts used in the text. Given their multimodal capabilities and free accessibility, many VLMs are often considered potential tools for font recognition. This raises a fundamental question: Do VLMs truly possess the capability to recognize fonts? To investigate this, we introduce the Font Recognition Benchmark (FRB), a compact and well-structured dataset comprising 15 commonly used fonts. FRB includes two versions: (i) an easy version, where 10 sentences are rendered in different fonts, and (ii) a hard version, where each text sample consists of the names of the 15 fonts themselves, introducing a stroop effect that challenges model perception. Through extensive evaluation of various VLMs on font recognition tasks, we arrive at the following key findings: (i) Current VLMs exhibit limited font recognition capabilities, with many state-of-the-art models failing to achieve satisfactory performance. (ii) Few-shot learning and Chain-of-Thought (CoT) prompting provide minimal benefits in improving font recognition accuracy across different VLMs. (iii) Attention analysis sheds light on the inherent limitations of VLMs in capturing semantic features.

☆ STI-Bench: Are MLLMs Ready for Precise Spatial-Temporal World Understanding?

Yun Li, Yiming Zhang, Tao Lin, XiangRui Liu, Wenxiao Cai, Zheng Liu, Bo Zhao

The use of Multimodal Large Language Models (MLLMs) as an end-to-end solution for Embodied AI and Autonomous Driving has become a prevailing trend. While MLLMs have been extensively studied for visual semantic understanding tasks, their ability to perform precise and quantitative spatial-temporal understanding in real-world applications remains largely unexamined, leading to uncertain prospects. To evaluate models' Spatial-Temporal Intelligence, we introduce STI-Bench, a benchmark designed to evaluate MLLMs' spatial-temporal understanding through challenging tasks such as estimating and predicting the appearance, pose, displacement, and motion of objects. Our benchmark encompasses a wide range of robot and vehicle operations across desktop, indoor, and outdoor scenarios. The extensive experiments reveals that the state-of-the-art MLLMs still struggle in real-world spatial-temporal understanding, especially in tasks requiring precise distance estimation and motion analysis.

☆ WaveFormer: A 3D Transformer with Wavelet-Driven Feature Representation for Efficient Medical Image Segmentation

Md Mahfuz Al Hasan, Mahdi Zaman, Abdul Jawad, Alberto Santamaria-Pang, Ho Hin Lee, Ivan Tarapov, Kyle See, Md Shah Imran, Antika Roy, Yaser Pourmohammadi Fallah, Navid Asadizanjani, Reza Forghani

Transformer-based architectures have advanced medical image analysis by effectively modeling long-range dependencies, yet they often struggle in 3D settings due to substantial memory overhead and insufficient capture of fine-grained local features. We address these limi- tations with WaveFormer, a novel 3D-transformer that: i) leverages the fundamental frequency-domain properties of features for contextual rep- resentation, and ii) is inspired by the top-down mechanism of the human visual recognition system, making it a biologically motivated architec- ture. By employing discrete wavelet transformations (DWT) at multiple scales, WaveFormer preserves both global context and high-frequency de- tails while replacing heavy upsampling layers with efficient wavelet-based summarization and reconstruction. This significantly reduces the number of parameters, which is critical for real-world deployment where compu- tational resources and training times are constrained. Furthermore, the model is generic and easily adaptable to diverse applications. Evaluations on BraTS2023, FLARE2021, and KiTS2023 demonstrate performance on par with state-of-the-art methods while offering substantially lower computational complexity.

☆ StrokeFusion: Vector Sketch Generation via Joint Stroke-UDF Encoding and Latent Sequence Diffusion

Jin Zhou, Yi Zhou, Pengfei Xu, Hui Huang

In the field of sketch generation, raster-format trained models often produce non-stroke artifacts, while vector-format trained models typically lack a holistic understanding of sketches, leading to compromised recognizability. Moreover, existing methods struggle to extract common features from similar elements (e.g., eyes of animals) appearing at varying positions across sketches. To address these challenges, we propose StrokeFusion, a two-stage framework for vector sketch generation. It contains a dual-modal sketch feature learning network that maps strokes into a high-quality latent space. This network decomposes sketches into normalized strokes and jointly encodes stroke sequences with Unsigned Distance Function (UDF) maps, representing sketches as sets of stroke feature vectors. Building upon this representation, our framework exploits a stroke-level latent diffusion model that simultaneously adjusts stroke position, scale, and trajectory during generation. This enables high-fidelity sketch generation while supporting stroke interpolation editing. Extensive experiments on the QuickDraw dataset demonstrate that our framework outperforms state-of-the-art techniques, validating its effectiveness in preserving structural integrity and semantic features. Code and models will be made publicly available upon publication.

☆ Decoupled Distillation to Erase: A General Unlearning Method for Any Class-centric Tasks CVPR2025

Yu Zhou, Dian Zheng, Qijie Mo, Renjie Lu, Kun-Yu Lin, Wei-Shi Zheng

In this work, we present DEcoupLEd Distillation To Erase (DELETE), a general and strong unlearning method for any class-centric tasks. To derive this, we first propose a theoretical framework to analyze the general form of unlearning loss and decompose it into forgetting and retention terms. Through the theoretical framework, we point out that a class of previous methods could be mainly formulated as a loss that implicitly optimizes the forgetting term while lacking supervision for the retention term, disturbing the distribution of pre-trained model and struggling to adequately preserve knowledge of the remaining classes. To address it, we refine the retention term using "dark knowledge" and propose a mask distillation unlearning method. By applying a mask to separate forgetting logits from retention logits, our approach optimizes both the forgetting and refined retention components simultaneously, retaining knowledge of the remaining classes while ensuring thorough forgetting of the target class. Without access to the remaining data or intervention (i.e., used in some works), we achieve state-of-the-art performance across various benchmarks. What's more, DELETE is a general solution that can be applied to various downstream tasks, including face recognition, backdoor defense, and semantic segmentation with great performance.

comment: CVPR2025, Equal contributions from first two authors

☆ Consistency-aware Self-Training for Iterative-based Stereo Matching CVPR 2025

Jingyi Zhou, Peng Ye, Haoyu Zhang, Jiakang Yuan, Rao Qiang, Liu YangChenXu, Wu Cailin, Feng Xu, Tao Chen

Iterative-based methods have become mainstream in stereo matching due to their high performance. However, these methods heavily rely on labeled data and face challenges with unlabeled real-world data. To this end, we propose a consistency-aware self-training framework for iterative-based stereo matching for the first time, leveraging real-world unlabeled data in a teacher-student manner. We first observe that regions with larger errors tend to exhibit more pronounced oscillation characteristics during model prediction.Based on this, we introduce a novel consistency-aware soft filtering module to evaluate the reliability of teacher-predicted pseudo-labels, which consists of a multi-resolution prediction consistency filter and an iterative prediction consistency filter to assess the prediction fluctuations of multiple resolutions and iterative optimization respectively. Further, we introduce a consistency-aware soft-weighted loss to adjust the weight of pseudo-labels accordingly, relieving the error accumulation and performance degradation problem due to incorrect pseudo-labels. Extensive experiments demonstrate that our method can improve the performance of various iterative-based stereo matching approaches in various scenarios. In particular, our method can achieve further enhancements over the current SOTA methods on several benchmark datasets.

comment: Accepted by CVPR 2025

☆ Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model

Dizhan Xue, Jing Cui, Shengsheng Qian, Chuanrui Hu, Changsheng Xu

Short-video platforms have gained immense popularity, captivating the interest of millions, if not billions, of users globally. Recently, researchers have highlighted the significance of analyzing the propagation of short-videos, which typically involves discovering commercial values, public opinions, user behaviors, etc. This paper proposes a new Short-video Propagation Influence Rating (SPIR) task and aims to promote SPIR from both the dataset and method perspectives. First, we propose a new Cross-platform Short-Video (XS-Video) dataset, which aims to provide a large-scale and real-world short-video propagation network across various platforms to facilitate the research on short-video propagation. Our XS-Video dataset includes 117,720 videos, 381,926 samples, and 535 topics across 5 biggest Chinese platforms, annotated with the propagation influence from level 0 to 9. To the best of our knowledge, this is the first large-scale short-video dataset that contains cross-platform data or provides all of the views, likes, shares, collects, fans, comments, and comment content. Second, we propose a Large Graph Model (LGM) named NetGPT, based on a novel three-stage training mechanism, to bridge heterogeneous graph-structured data with the powerful reasoning ability and knowledge of Large Language Models (LLMs). Our NetGPT can comprehend and analyze the short-video propagation graph, enabling it to predict the long-term propagation influence of short-videos. Comprehensive experimental results evaluated by both classification and regression metrics on our XS-Video dataset indicate the superiority of our method for SPIR.

☆ Every Painting Awakened: A Training-free Framework for Painting-to-Animation Generation

Lingyu Liu, Yaxiong Wang, Li Zhu, Zhedong Zheng

We introduce a training-free framework specifically designed to bring real-world static paintings to life through image-to-video (I2V) synthesis, addressing the persistent challenge of aligning these motions with textual guidance while preserving fidelity to the original artworks. Existing I2V methods, primarily trained on natural video datasets, often struggle to generate dynamic outputs from static paintings. It remains challenging to generate motion while maintaining visual consistency with real-world paintings. This results in two distinct failure modes: either static outputs due to limited text-based motion interpretation or distorted dynamics caused by inadequate alignment with real-world artistic styles. We leverage the advanced text-image alignment capabilities of pre-trained image models to guide the animation process. Our approach introduces synthetic proxy images through two key innovations: (1) Dual-path score distillation: We employ a dual-path architecture to distill motion priors from both real and synthetic data, preserving static details from the original painting while learning dynamic characteristics from synthetic frames. (2) Hybrid latent fusion: We integrate hybrid features extracted from real paintings and synthetic proxy images via spherical linear interpolation in the latent space, ensuring smooth transitions and enhancing temporal consistency. Experimental evaluations confirm that our approach significantly improves semantic alignment with text prompts while faithfully preserving the unique characteristics and integrity of the original paintings. Crucially, by achieving enhanced dynamic effects without requiring any model training or learnable parameters, our framework enables plug-and-play integration with existing I2V methods, making it an ideal solution for animating real-world paintings. More animated examples can be found on our project website.

comment: The project is available at: https://painting-animation.github.io/animation/

☆ AdaMMS: Model Merging for Heterogeneous Multimodal Large Language Models with Unsupervised Coefficient Optimization CVPR 2025

Yiyang Du, Xiaochen Wang, Chi Chen, Jiabo Ye, Yiru Wang, Peng Li, Ming Yan, Ji Zhang, Fei Huang, Zhifang Sui, Maosong Sun, Yang Liu

Recently, model merging methods have demonstrated powerful strengths in combining abilities on various tasks from multiple Large Language Models (LLMs). While previous model merging methods mainly focus on merging homogeneous models with identical architecture, they meet challenges when dealing with Multimodal Large Language Models (MLLMs) with inherent heterogeneous property, including differences in model architecture and the asymmetry in the parameter space. In this work, we propose AdaMMS, a novel model merging method tailored for heterogeneous MLLMs. Our method tackles the challenges in three steps: mapping, merging and searching. Specifically, we first design mapping function between models to apply model merging on MLLMs with different architecture. Then we apply linear interpolation on model weights to actively adapt the asymmetry in the heterogeneous MLLMs. Finally in the hyper-parameter searching step, we propose an unsupervised hyper-parameter selection method for model merging. As the first model merging method capable of merging heterogeneous MLLMs without labeled data, extensive experiments on various model combinations demonstrated that AdaMMS outperforms previous model merging methods on various vision-language benchmarks.

comment: CVPR 2025

☆ Investigation of intelligent barbell squat coaching system based on computer vision and machine learning

Yinq-Rong Chern, Yuhao Lee, Hsiao-Ching Lin, Guan-Ting Chen, Ying-Hsien Chen, Fu-Sung Lin, Chih-Yao Chuang, Jenn-Jier James Lien, Chih-Hsien Huang

Purpose: Research has revealed that strength training can reduce the incidence of chronic diseases and physical deterioration at any age. Therefore, having a movement diagnostic system is crucial for training alone. Hence, this study developed an artificial intelligence and computer vision-based barbell squat coaching system with a real-time mode that immediately diagnoses the issue and provides feedback after each squat. In addition, a replay mode allows users to examine their previous squats and check their comments. Initially, four primary characteristics of the barbell squat were identified: body joint angles, dorsiflexion, the ratio of knee-to-hip movement, and barbell stability. Methods: We collect 8,151 squats from 77 participants, categorizing them as good squats and six issues. Then, we trained the diagnosis models with three machine-learning architectures. Furthermore, this research applied the SHapley Additive exPlanations (SHAP) method to enhance the accuracy of issue prediction and reduce the computation time by feature selection. Results: The F1 score of the six issues reached 86.86%, 69.01%, 77.42%, 90.74%, 95.83%, and 100%. Each squat diagnosis took less than 0.5 seconds. Finally, this study examined the efficacy of the proposed system with two groups of participants trained with and without the system. Subsequently, participants trained with the system exhibited substantial improvements in their squat technique, as assessed both by the system itself and by a professional weightlifting coach. Conclusion: This is a comprehensive study that integrates artificial intelligence, computer vision and multivariable processing technologies, aimed at building a real-time, user-friendly barbell squat feedback and training system.

☆ KOFFVQA: An Objectively Evaluated Free-form VQA Benchmark for Large Vision-Language Models in the Korean Language CVPR

Yoonshik Kim, Jaeyoon Jung

The recent emergence of Large Vision-Language Models(VLMs) has resulted in a variety of different benchmarks for evaluating such models. Despite this, we observe that most existing evaluation methods suffer from the fact that they either require the model to choose from pre-determined responses, sacrificing open-endedness, or evaluate responses using a judge model, resulting in subjective and unreliable evaluation. In addition, we observe a lack of benchmarks for VLMs in the Korean language, which are necessary as a separate metric from more common English language benchmarks, as the performance of generative language models can differ significantly based on the language being used. Therefore, we present KOFFVQA, a general-purpose free-form visual question answering benchmark in the Korean language for the evaluation of VLMs. Our benchmark consists of 275 carefully crafted questions each paired with an image and grading criteria covering 10 different aspects of VLM performance. The grading criteria eliminate the problem of unreliability by allowing the judge model to grade each response based on a pre-determined set of rules. By defining the evaluation criteria in an objective manner, even a small open-source model can be used to evaluate models on our benchmark reliably. In addition to evaluating a large number of existing VLMs on our benchmark, we also experimentally verify that our method of using pre-existing grading criteria for evaluation is much more reliable than existing methods. Our evaluation code is available at https://github.com/maum-ai/KOFFVQA

comment: Accepted to CVPRW 2025, Workshop on Benchmarking and Expanding AI Multimodal Approaches

☆ Exploring Temporal Dynamics in Event-based Eye Tracker CVPR 2025

Hongwei Ren, Xiaopeng Lin, Hongxiang Huang, Yue Zhou, Bojun Cheng

Eye-tracking is a vital technology for human-computer interaction, especially in wearable devices such as AR, VR, and XR. The realization of high-speed and high-precision eye-tracking using frame-based image sensors is constrained by their limited temporal resolution, which impairs the accurate capture of rapid ocular dynamics, such as saccades and blinks. Event cameras, inspired by biological vision systems, are capable of perceiving eye movements with extremely low power consumption and ultra-high temporal resolution. This makes them a promising solution for achieving high-speed, high-precision tracking with rich temporal dynamics. In this paper, we propose TDTracker, an effective eye-tracking framework that captures rapid eye movements by thoroughly modeling temporal dynamics from both implicit and explicit perspectives. TDTracker utilizes 3D convolutional neural networks to capture implicit short-term temporal dynamics and employs a cascaded structure consisting of a Frequency-aware Module, GRU, and Mamba to extract explicit long-term temporal dynamics. Ultimately, a prediction heatmap is used for eye coordinate regression. Experimental results demonstrate that TDTracker achieves state-of-the-art (SOTA) performance on the synthetic SEET dataset and secured Third place in the CVPR event-based eye-tracking challenge 2025. Our code is available at https://github.com/rhwxmx/TDTracker.

comment: Accepted by CVPR 2025 Event-based Vision Workshop

☆ LATex: Leveraging Attribute-based Text Knowledge for Aerial-Ground Person Re-Identification

Xiang Hu, Yuhao Wang, Pingping Zhang, Huchuan Lu

Aerial-Ground person Re-IDentification (AG-ReID) aims to retrieve specific persons across heterogeneous cameras in different views. Previous methods usually adopt large-scale models, focusing on view-invariant features. However, they overlook the semantic information in person attributes. Additionally, existing training strategies often rely on full fine-tuning large-scale models, which significantly increases training costs. To address these issues, we propose a novel framework named LATex for AG-ReID, which adopts prompt-tuning strategies to leverage attribute-based text knowledge. More specifically, we first introduce the Contrastive Language-Image Pre-training (CLIP) model as the backbone, and propose an Attribute-aware Image Encoder (AIE) to extract global semantic features and attribute-aware features. Then, with these features, we propose a Prompted Attribute Classifier Group (PACG) to generate person attribute predictions and obtain the encoded representations of predicted attributes. Finally, we design a Coupled Prompt Template (CPT) to transform attribute tokens and view information into structured sentences. These sentences are processed by the text encoder of CLIP to generate more discriminative features. As a result, our framework can fully leverage attribute-based text knowledge to improve the AG-ReID. Extensive experiments on three AG-ReID benchmarks demonstrate the effectiveness of our proposed LATex. The source code will be available.

☆ Effective Cloud Removal for Remote Sensing Images by an Improved Mean-Reverting Denoising Model with Elucidated Design Space

Yi Liu, Wengen Li, Jihong Guan, Shuigeng Zhou, Yichao Zhang

Cloud removal (CR) remains a challenging task in remote sensing image processing. Although diffusion models (DM) exhibit strong generative capabilities, their direct applications to CR are suboptimal, as they generate cloudless images from random noise, ignoring inherent information in cloudy inputs. To overcome this drawback, we develop a new CR model EMRDM based on mean-reverting diffusion models (MRDMs) to establish a direct diffusion process between cloudy and cloudless images. Compared to current MRDMs, EMRDM offers a modular framework with updatable modules and an elucidated design space, based on a reformulated forward process and a new ordinary differential equation (ODE)-based backward process. Leveraging our framework, we redesign key MRDM modules to boost CR performance, including restructuring the denoiser via a preconditioning technique, reorganizing the training process, and improving the sampling process by introducing deterministic and stochastic samplers. To achieve multi-temporal CR, we further develop a denoising network for simultaneously denoising sequential images. Experiments on mono-temporal and multi-temporal datasets demonstrate the superior performance of EMRDM. Our code is available at https://github.com/Ly403/EMRDM.

comment: 29 pages, 12 figures

☆ HOIGen-1M: A Large-scale Dataset for Human-Object Interaction Video Generation CVPR 2025

Kun Liu, Qi Liu, Xinchen Liu, Jie Li, Yongdong Zhang, Jiebo Luo, Xiaodong He, Wu Liu

Text-to-video (T2V) generation has made tremendous progress in generating complicated scenes based on texts. However, human-object interaction (HOI) often cannot be precisely generated by current T2V models due to the lack of large-scale videos with accurate captions for HOI. To address this issue, we introduce HOIGen-1M, the first largescale dataset for HOI Generation, consisting of over one million high-quality videos collected from diverse sources. In particular, to guarantee the high quality of videos, we first design an efficient framework to automatically curate HOI videos using the powerful multimodal large language models (MLLMs), and then the videos are further cleaned by human annotators. Moreover, to obtain accurate textual captions for HOI videos, we design a novel video description method based on a Mixture-of-Multimodal-Experts (MoME) strategy that not only generates expressive captions but also eliminates the hallucination by individual MLLM. Furthermore, due to the lack of an evaluation framework for generated HOI videos, we propose two new metrics to assess the quality of generated videos in a coarse-to-fine manner. Extensive experiments reveal that current T2V models struggle to generate high-quality HOI videos and confirm that our HOIGen-1M dataset is instrumental for improving HOI video generation. Project webpage is available at https://liuqi-creat.github.io/HOIGen.github.io.

comment: CVPR 2025

☆ ElimPCL: Eliminating Noise Accumulation with Progressive Curriculum Labeling for Source-Free Domain Adaptation ICME 2025

Jie Cheng, Hao Zheng, Meiguang Zheng, Lei Wang, Hao Wu, Jian Zhang

Source-Free Domain Adaptation (SFDA) aims to train a target model without source data, and the key is to generate pseudo-labels using a pre-trained source model. However, we observe that the source model often produces highly uncertain pseudo-labels for hard samples, particularly those heavily affected by domain shifts, leading to these noisy pseudo-labels being introduced even before adaptation and further reinforced through parameter updates. Additionally, they continuously influence neighbor samples through propagation in the feature space.To eliminate the issue of noise accumulation, we propose a novel Progressive Curriculum Labeling (ElimPCL) method, which iteratively filters trustworthy pseudo-labeled samples based on prototype consistency to exclude high-noise samples from training. Furthermore, a Dual MixUP technique is designed in the feature space to enhance the separability of hard samples, thereby mitigating the interference of noisy samples on their neighbors.Extensive experiments validate the effectiveness of ElimPCL, achieving up to a 3.4% improvement on challenging tasks compared to state-of-the-art methods.

comment: ICME 2025 camera-ready

☆ Expanding-and-Shrinking Binary Neural Networks

Xulong Shi, Caiyi Sun, Zhi Qi, Liu Hao, Xiaodong Yang

While binary neural networks (BNNs) offer significant benefits in terms of speed, memory and energy, they encounter substantial accuracy degradation in challenging tasks compared to their real-valued counterparts. Due to the binarization of weights and activations, the possible values of each entry in the feature maps generated by BNNs are strongly constrained. To tackle this limitation, we propose the expanding-and-shrinking operation, which enhances binary feature maps with negligible increase of computation complexity, thereby strengthening the representation capacity. Extensive experiments conducted on multiple benchmarks reveal that our approach generalizes well across diverse applications ranging from image classification, object detection to generative diffusion model, while also achieving remarkable improvement over various leading binarization algorithms based on different architectures including both CNNs and Transformers.

☆ 3D Dental Model Segmentation with Geometrical Boundary Preserving

Shufan Xi, Zexian Liu, Junlin Chang, Hongyu Wu, Xiaogang Wang, Aimin Hao

3D intraoral scan mesh is widely used in digital dentistry diagnosis, segmenting 3D intraoral scan mesh is a critical preliminary task. Numerous approaches have been devised for precise tooth segmentation. Currently, the deep learning-based methods are capable of the high accuracy segmentation of crown. However, the segmentation accuracy at the junction between the crown and the gum is still below average. Existing down-sampling methods are unable to effectively preserve the geometric details at the junction. To address these problems, we propose CrossTooth, a boundary-preserving segmentation method that combines 3D mesh selective downsampling to retain more vertices at the tooth-gingiva area, along with cross-modal discriminative boundary features extracted from multi-view rendered images, enhancing the geometric representation of the segmentation network. Using a point network as a backbone and incorporating image complementary features, CrossTooth significantly improves segmentation accuracy, as demonstrated by experiments on a public intraoral scan dataset.

comment: The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025

☆ Detail-aware multi-view stereo network for depth estimation

Haitao Tian, Junyang Li, Chenxing Wang, Helong Jiang

Multi-view stereo methods have achieved great success for depth estimation based on the coarse-to-fine depth learning frameworks, however, the existing methods perform poorly in recovering the depth of object boundaries and detail regions. To address these issues, we propose a detail-aware multi-view stereo network (DA-MVSNet) with a coarse-to-fine framework. The geometric depth clues hidden in the coarse stage are utilized to maintain the geometric structural relationships between object surfaces and enhance the expressive capability of image features. In addition, an image synthesis loss is employed to constrain the gradient flow for detailed regions and further strengthen the supervision of object boundaries and texture-rich areas. Finally, we propose an adaptive depth interval adjustment strategy to improve the accuracy of object reconstruction. Extensive experiments on the DTU and Tanks & Temples datasets demonstrate that our method achieves competitive results. The code is available at https://github.com/wsmtht520-/DAMVSNet.

☆ The Devil is in the Distributions: Explicit Modeling of Scene Content is Key in Zero-Shot Video Captioning

Mingkai Tian, Guorong Li, Yuankai Qi, Amin Beheshti, Javen Qinfeng Shi, Anton van den Hengel, Qingming Huang

Zero-shot video captioning requires that a model generate high-quality captions without human-annotated video-text pairs for training. State-of-the-art approaches to the problem leverage CLIP to extract visual-relevant textual prompts to guide language models in generating captions. These methods tend to focus on one key aspect of the scene and build a caption that ignores the rest of the visual input. To address this issue, and generate more accurate and complete captions, we propose a novel progressive multi-granularity textual prompting strategy for zero-shot video captioning. Our approach constructs three distinct memory banks, encompassing noun phrases, scene graphs of noun phrases, and entire sentences. Moreover, we introduce a category-aware retrieval mechanism that models the distribution of natural language surrounding the specific topics in question. Extensive experiments demonstrate the effectiveness of our method with 5.7%, 16.2%, and 3.4% improvements in terms of the main metric CIDEr on MSR-VTT, MSVD, and VATEX benchmarks compared to existing state-of-the-art.

comment: 13 pages

☆ Learning Bijective Surface Parameterization for Inferring Signed Distance Functions from Sparse Point Clouds with Grid Deformation CVPR

Takeshi Noda, Chao Chen, Junsheng Zhou, Weiqi Zhang, Yu-Shen Liu, Zhizhong Han

Inferring signed distance functions (SDFs) from sparse point clouds remains a challenge in surface reconstruction. The key lies in the lack of detailed geometric information in sparse point clouds, which is essential for learning a continuous field. To resolve this issue, we present a novel approach that learns a dynamic deformation network to predict SDFs in an end-to-end manner. To parameterize a continuous surface from sparse points, we propose a bijective surface parameterization (BSP) that learns the global shape from local patches. Specifically, we construct a bijective mapping for sparse points from the parametric domain to 3D local patches, integrating patches into the global surface. Meanwhile, we introduce grid deformation optimization (GDO) into the surface approximation to optimize the deformation of grid points and further refine the parametric surfaces. Experimental results on synthetic and real scanned datasets demonstrate that our method significantly outperforms the current state-of-the-art methods. Project page: https://takeshie.github.io/Bijective-SDF

comment: Accepted by Conference on Computer Vision and Pattern Recognition (CVPR) 2025. Project page:https://takeshie.github.io/Bijective-SDF

☆ Context-Independent OCR with Multimodal LLMs: Effects of Image Resolution and Visual Complexity

Kotaro Inoue

Due to their high versatility in tasks such as image captioning, document analysis, and automated content generation, multimodal Large Language Models (LLMs) have attracted significant attention across various industrial fields. In particular, they have been shown to surpass specialized models in Optical Character Recognition (OCR). Nevertheless, their performance under different image conditions remains insufficiently investigated, and individual character recognition is not guaranteed due to their reliance on contextual cues. In this work, we examine a context-independent OCR task using single-character images with diverse visual complexities to determine the conditions for accurate recognition. Our findings reveal that multimodal LLMs can match conventional OCR methods at about 300 ppi, yet their performance deteriorates significantly below 150 ppi. Additionally, we observe a very weak correlation between visual complexity and misrecognitions, whereas a conventional OCR-specific model exhibits no correlation. These results suggest that image resolution and visual complexity may play an important role in the reliable application of multimodal LLMs to OCR tasks that require precise character-level accuracy.

☆ LiM-Loc: Visual Localization with Dense and Accurate 3D Reference Maps Directly Corresponding 2D Keypoints to 3D LiDAR Point Clouds

Masahiko Tsuji, Hitoshi Niigaki, Ryuichi Tanida

Visual localization is to estimate the 6-DOF camera pose of a query image in a 3D reference map. We extract keypoints from the reference image and generate a 3D reference map with 3D reconstruction of the keypoints in advance. We emphasize that the more keypoints in the 3D reference map and the smaller the error of the 3D positions of the keypoints, the higher the accuracy of the camera pose estimation. However, previous image-only methods require a huge number of images, and it is difficult to 3D-reconstruct keypoints without error due to inevitable mismatches and failures in feature matching. As a result, the 3D reference map is sparse and inaccurate. In contrast, accurate 3D reference maps can be generated by combining images and 3D sensors. Recently, 3D-LiDAR has been widely used around the world. LiDAR, which measures a large space with high density, has become inexpensive. In addition, accurately calibrated cameras are also widely used, so images that record the external parameters of the camera without errors can be easily obtained. In this paper, we propose a method to directly assign 3D LiDAR point clouds to keypoints to generate dense and accurate 3D reference maps. The proposed method avoids feature matching and achieves accurate 3D reconstruction for almost all keypoints. To estimate camera pose over a wide area, we use the wide-area LiDAR point cloud to remove points that are not visible to the camera and reduce 2D-3D correspondence errors. Using indoor and outdoor datasets, we apply the proposed method to several state-of-the-art local features and confirm that it improves the accuracy of camera pose estimation.

comment: 8 pages, 6 figures

☆ DeepDubber-V1: Towards High Quality and Dialogue, Narration, Monologue Adaptive Movie Dubbing Via Multi-Modal Chain-of-Thoughts Reasoning Guidance

Junjie Zheng, Zihao Chen, Chaofan Ding, Xinhan Di

Current movie dubbing technology can generate the desired voice from a given speech prompt, ensuring good synchronization between speech and visuals while accurately conveying the intended emotions. However, in movie dubbing, key aspects such as adapting to different dubbing styles, handling dialogue, narration, and monologue effectively, and understanding subtle details like the age and gender of speakers, have not been well studied. To address this challenge, we propose a framework of multi-modal large language model. First, it utilizes multimodal Chain-of-Thought (CoT) reasoning methods on visual inputs to understand dubbing styles and fine-grained attributes. Second, it generates high-quality dubbing through large speech generation models, guided by multimodal conditions. Additionally, we have developed a movie dubbing dataset with CoT annotations. The evaluation results demonstrate a performance improvement over state-of-the-art methods across multiple datasets. In particular, for the evaluation metrics, the SPK-SIM and EMO-SIM increases from 82.48% to 89.74%, 66.24% to 78.88% for dubbing setting 2.0 on V2C Animation dataset, LSE-D and MCD-SL decreases from 14.79 to 14.63, 5.24 to 4.74 for dubbing setting 2.0 on Grid dataset, SPK-SIM increases from 64.03 to 83.42 and WER decreases from 52.69% to 23.20% for initial reasoning setting on proposed CoT-Movie-Dubbing dataset in the comparison with the state-of-the art models.

comment: 11 pages, 5 figures

☆ Introducing the Short-Time Fourier Kolmogorov Arnold Network: A Dynamic Graph CNN Approach for Tree Species Classification in 3D Point Clouds

Said Ohamouddoua, Mohamed Ohamouddoub, Rafik Lasrib, Hanaa El Afiaa, Raddouane Chiheba, Abdellatif El Afiaa

Accurate classification of tree species based on Terrestrial Laser Scanning (TLS) and Airborne Laser Scanning (ALS) is essential for biodiversity conservation. While advanced deep learning models for 3D point cloud classification have demonstrated strong performance in this domain, their high complexity often hinders the development of efficient, low-computation architectures. In this paper, we introduce STFT-KAN, a novel Kolmogorov-Arnold network that integrates the Short-Time Fourier Transform (STFT), which can replace the standard linear layer with activation. We implemented STFT-KAN within a lightweight version of DGCNN, called liteDGCNN, to classify tree species using the TLS data. Our experiments show that STFT-KAN outperforms existing KAN variants by effectively balancing model complexity and performance with parameter count reduction, achieving competitive results compared to MLP-based models. Additionally, we evaluated a hybrid architecture that combines MLP in edge convolution with STFT-KAN in other layers, achieving comparable performance to MLP models while reducing the parameter count by 50% and 75% compared to other KAN-based variants. Furthermore, we compared our model to leading 3D point cloud learning approaches, demonstrating that STFT-KAN delivers competitive results compared to the state-of-the-art method PointMLP lite with an 87% reduction in parameter count.

☆ Uni-Render: A Unified Accelerator for Real-Time Rendering Across Diverse Neural Renderers HPCA'25

Chaojian Li, Sixu Li, Linrui Jiang, Jingqun Zhang, Yingyan Celine Lin

Recent advancements in neural rendering technologies and their supporting devices have paved the way for immersive 3D experiences, significantly transforming human interaction with intelligent devices across diverse applications. However, achieving the desired real-time rendering speeds for immersive interactions is still hindered by (1) the lack of a universal algorithmic solution for different application scenarios and (2) the dedication of existing devices or accelerators to merely specific rendering pipelines. To overcome this challenge, we have developed a unified neural rendering accelerator that caters to a wide array of typical neural rendering pipelines, enabling real-time and on-device rendering across different applications while maintaining both efficiency and compatibility. Our accelerator design is based on the insight that, although neural rendering pipelines vary and their algorithm designs are continually evolving, they typically share common operators, predominantly executing similar workloads. Building on this insight, we propose a reconfigurable hardware architecture that can dynamically adjust dataflow to align with specific rendering metric requirements for diverse applications, effectively supporting both typical and the latest hybrid rendering pipelines. Benchmarking experiments and ablation studies on both synthetic and real-world scenes demonstrate the effectiveness of the proposed accelerator. The proposed unified accelerator stands out as the first solution capable of achieving real-time neural rendering across varied representative pipelines on edge devices, potentially paving the way for the next generation of neural graphics applications.

comment: Accepted by HPCA'25

☆ RIG: Synergizing Reasoning and Imagination in End-to-End Generalist Policy

Zhonghan Zhao, Wenwei Zhang, Haian Huang, Kuikun Liu, Jianfei Gao, Gaoang Wang, Kai Chen

Reasoning before action and imagining potential outcomes (i.e., world models) are essential for embodied agents operating in complex open-world environments. Yet, prior work either incorporates only one of these abilities in an end-to-end agent or integrates multiple specialized models into an agent system, limiting the learning efficiency and generalization of the policy. Thus, this paper makes the first attempt to synergize Reasoning and Imagination in an end-to-end Generalist policy, termed RIG. To train RIG in an end-to-end manner, we construct a data pipeline that progressively integrates and enriches the content of imagination and reasoning in the trajectories collected from existing agents. The joint learning of reasoning and next image generation explicitly models the inherent correlation between reasoning, action, and dynamics of environments, and thus exhibits more than $17\times$ sample efficiency improvements and generalization in comparison with previous works. During inference, RIG first reasons about the next action, produces potential action, and then predicts the action outcomes, which offers the agent a chance to review and self-correct based on the imagination before taking real actions. Experimental results show that the synergy of reasoning and imagination not only improves the robustness, generalization, and interoperability of generalist policy but also enables test-time scaling to enhance overall performance.

♻ ☆ Image as an IMU: Estimating Camera Motion from a Single Motion-Blurred Image

Jerred Chen, Ronald Clark

In many robotics and VR/AR applications, fast camera motions cause a high level of motion blur, causing existing camera pose estimation methods to fail. In this work, we propose a novel framework that leverages motion blur as a rich cue for motion estimation rather than treating it as an unwanted artifact. Our approach works by predicting a dense motion flow field and a monocular depth map directly from a single motion-blurred image. We then recover the instantaneous camera velocity by solving a linear least squares problem under the small motion assumption. In essence, our method produces an IMU-like measurement that robustly captures fast and aggressive camera movements. To train our model, we construct a large-scale dataset with realistic synthetic motion blur derived from ScanNet++v2 and further refine our model by training end-to-end on real data using our fully differentiable pipeline. Extensive evaluations on real-world benchmarks demonstrate that our method achieves state-of-the-art angular and translational velocity estimates, outperforming current methods like MASt3R and COLMAP.

comment: Project page: https://jerredchen.github.io/image-as-imu/

♻ ☆ A Double Deep Learning-based Solution for Efficient Event Data Coding and Classification

Abdelrahman Seleem, André F. R. Guarda, Nuno M. M. Rodrigues, Fernando Pereira

Event cameras have the ability to capture asynchronous per-pixel brightness changes, called "events", offering advantages over traditional frame-based cameras for computer vision applications. Efficiently coding event data is critical for transmission and storage, given the significant volume of events. This paper proposes a novel double deep learning-based architecture for both event data coding and classification, using a point cloud-based representation for events. In this context, the conversions from events to point clouds and back to events are key steps in the proposed solution, and therefore its impact is evaluated in terms of compression and classification performance. Experimental results show that it is possible to achieve a classification performance of compressed events which is similar to one of the original events, even after applying a lossy point cloud codec, notably the recent learning-based JPEG Pleno Point Cloud Coding standard, with a clear rate reduction. Experimental results also demonstrate that events coded using JPEG PCC achieve better classification performance than those coded using the conventional lossy MPEG Geometry-based Point Cloud Coding standard. Furthermore, the adoption of learning-based coding offers high potential for performing computer vision tasks in the compressed domain, which allows skipping the decoding stage while mitigating the impact of coding artifacts.

♻ ☆ Reversible Decoupling Network for Single Image Reflection Removal CVPR 2025

Hao Zhao, Mingjia Li, Qiming Hu, Xiaojie Guo

Recent deep-learning-based approaches to single-image reflection removal have shown promising advances, primarily for two reasons: 1) the utilization of recognition-pretrained features as inputs, and 2) the design of dual-stream interaction networks. However, according to the Information Bottleneck principle, high-level semantic clues tend to be compressed or discarded during layer-by-layer propagation. Additionally, interactions in dual-stream networks follow a fixed pattern across different layers, limiting overall performance. To address these limitations, we propose a novel architecture called Reversible Decoupling Network (RDNet), which employs a reversible encoder to secure valuable information while flexibly decoupling transmission- and reflection-relevant features during the forward pass. Furthermore, we customize a transmission-rate-aware prompt generator to dynamically calibrate features, further boosting performance. Extensive experiments demonstrate the superiority of RDNet over existing SOTA methods on five widely-adopted benchmark datasets. RDNet achieves the best performance in the NTIRE 2025 Single Image Reflection Removal in the Wild Challenge in both fidelity and perceptual comparison. Our code is available at https://github.com/lime-j/RDNet

comment: To appear at CVPR 2025

♻ ☆ Perceptually Accurate 3D Talking Head Generation: New Definitions, Speech-Mesh Representation, and Evaluation Metrics CVPR 2025

Lee Chae-Yeon, Oh Hyun-Bin, Han EunGi, Kim Sung-Bin, Suekyeong Nam, Tae-Hyun Oh

Recent advancements in speech-driven 3D talking head generation have made significant progress in lip synchronization. However, existing models still struggle to capture the perceptual alignment between varying speech characteristics and corresponding lip movements. In this work, we claim that three criteria -- Temporal Synchronization, Lip Readability, and Expressiveness -- are crucial for achieving perceptually accurate lip movements. Motivated by our hypothesis that a desirable representation space exists to meet these three criteria, we introduce a speech-mesh synchronized representation that captures intricate correspondences between speech signals and 3D face meshes. We found that our learned representation exhibits desirable characteristics, and we plug it into existing models as a perceptual loss to better align lip movements to the given speech. In addition, we utilize this representation as a perceptual metric and introduce two other physically grounded lip synchronization metrics to assess how well the generated 3D talking heads align with these three criteria. Experiments show that training 3D talking head generation models with our perceptual loss significantly improve all three aspects of perceptually accurate lip synchronization. Codes and datasets are available at https://perceptual-3d-talking-head.github.io/.

comment: CVPR 2025. Project page: https://perceptual-3d-talking-head.github.io/

♻ ☆ The impact of internal variability on benchmarking deep learning climate emulators

Björn Lütjens, Raffaele Ferrari, Duncan Watson-Parris, Noelle Selin

Full-complexity Earth system models (ESMs) are computationally very expensive, limiting their use in exploring the climate outcomes of multiple emission pathways. More efficient emulators that approximate ESMs can directly map emissions onto climate outcomes, and benchmarks are being used to evaluate their accuracy on standardized tasks and datasets. We investigate a popular benchmark in data-driven climate emulation, ClimateBench, on which deep learning-based emulators are currently achieving the best performance. We compare these deep learning emulators with a linear regression-based emulator, akin to pattern scaling, and show that it outperforms the incumbent 100M-parameter deep learning foundation model, ClimaX, on 3 out of 4 regionally-resolved climate variables, notably surface temperature and precipitation. While emulating surface temperature is expected to be predominantly linear, this result is surprising for emulating precipitation. Precipitation is a much more noisy variable, and we show that deep learning emulators can overfit to internal variability noise at low frequencies, degrading their performance in comparison to a linear emulator. We address the issue of overfitting by increasing the number of climate simulations per emission pathway (from 3 to 50) and updating the benchmark targets with the respective ensemble averages from the MPI-ESM1.2-LR model. Using the new targets, we show that linear pattern scaling continues to be more accurate on temperature, but can be outperformed by a deep learning-based technique for emulating precipitation. We publish our code and data at github.com/blutjens/climate-emulator.

♻ ☆ CASA: Class-Agnostic Shared Attributes in Vision-Language Models for Efficient Incremental Object Detection

Mingyi Guo, Yuyang Liu, Zhiyuan Yan, Zongying Lin, Peixi Peng, Yonghong Tian

Incremental object detection is fundamentally challenged by catastrophic forgetting. A major factor contributing to this issue is background shift, where background categories in sequential tasks may overlap with either previously learned or future unseen classes. To address this, we propose a novel method called Class-Agnostic Shared Attribute Base (CASA) that encourages the model to learn category-agnostic attributes shared across incremental classes. Our approach leverages an LLM to generate candidate textual attributes, selects the most relevant ones based on the current training data, and records their importance in an assignment matrix. For subsequent tasks, the retained attributes are frozen, and new attributes are selected from the remaining candidates, ensuring both knowledge retention and adaptability. Extensive experiments on the COCO dataset demonstrate the state-of-the-art performance of our method.

♻ ☆ Finer-CAM: Spotting the Difference Reveals Finer Details for Visual Explanation CVPR 2025

Ziheng Zhang, Jianyang Gu, Arpita Chowdhury, Zheda Mai, David Carlyn, Tanya Berger-Wolf, Yu Su, Wei-Lun Chao

Class activation map (CAM) has been widely used to highlight image regions that contribute to class predictions. Despite its simplicity and computational efficiency, CAM often struggles to identify discriminative regions that distinguish visually similar fine-grained classes. Prior efforts address this limitation by introducing more sophisticated explanation processes, but at the cost of extra complexity. In this paper, we propose Finer-CAM, a method that retains CAM's efficiency while achieving precise localization of discriminative regions. Our key insight is that the deficiency of CAM lies not in "how" it explains, but in "what" it explains. Specifically, previous methods attempt to identify all cues contributing to the target class's logit value, which inadvertently also activates regions predictive of visually similar classes. By explicitly comparing the target class with similar classes and spotting their differences, Finer-CAM suppresses features shared with other classes and emphasizes the unique, discriminative details of the target class. Finer-CAM is easy to implement, compatible with various CAM methods, and can be extended to multi-modal models for accurate localization of specific concepts. Additionally, Finer-CAM allows adjustable comparison strength, enabling users to selectively highlight coarse object contours or fine discriminative details. Quantitatively, we show that masking out the top 5% of activated pixels by Finer-CAM results in a larger relative confidence drop compared to baselines. The source code and demo are available at https://github.com/Imageomics/Finer-CAM.

comment: Accepted by CVPR 2025

♻ ☆ DICE: Discrete Inversion Enabling Controllable Editing for Multinomial Diffusion and Masked Generative Models CVPR 2025

Xiaoxiao He, Ligong Han, Quan Dao, Song Wen, Minhao Bai, Di Liu, Han Zhang, Martin Renqiang Min, Felix Juefei-Xu, Chaowei Tan, Bo Liu, Kang Li, Hongdong Li, Junzhou Huang, Faez Ahmed, Akash Srivastava, Dimitris Metaxas

Discrete diffusion models have achieved success in tasks like image generation and masked language modeling but face limitations in controlled content editing. We introduce DICE (Discrete Inversion for Controllable Editing), the first approach to enable precise inversion for discrete diffusion models, including multinomial diffusion and masked generative models. By recording noise sequences and masking patterns during the reverse diffusion process, DICE enables accurate reconstruction and flexible editing of discrete data without the need for predefined masks or attention manipulation. We demonstrate the effectiveness of DICE across both image and text domains, evaluating it on models such as VQ-Diffusion, Paella, and RoBERTa. Our results show that DICE preserves high data fidelity while enhancing editing capabilities, offering new opportunities for fine-grained content manipulation in discrete spaces.

comment: Project webpage: https://hexiaoxiao-cs.github.io/DICE/. This paper was accepted to CVPR 2025 but later desk-rejected post camera-ready, due to a withdrawal from ICLR made 14 days before reviewer assignment

♻ ☆ Gen3DSR: Generalizable 3D Scene Reconstruction via Divide and Conquer from a Single View 3DV 2025

Andreea Ardelean, Mert Özer, Bernhard Egger

Single-view 3D reconstruction is currently approached from two dominant perspectives: reconstruction of scenes with limited diversity using 3D data supervision or reconstruction of diverse singular objects using large image priors. However, real-world scenarios are far more complex and exceed the capabilities of these methods. We therefore propose a hybrid method following a divide-and-conquer strategy. We first process the scene holistically, extracting depth and semantic information, and then leverage an object-level method for the detailed reconstruction of individual components. By splitting the problem into simpler tasks, our system is able to generalize to various types of scenes without retraining or fine-tuning. We purposely design our pipeline to be highly modular with independent, self-contained modules, to avoid the need for end-to-end training of the whole system. This enables the pipeline to naturally improve as future methods can replace the individual modules. We demonstrate the reconstruction performance of our approach on both synthetic and real-world scenes, comparing favorable against prior works. Project page: https://andreeadogaru.github.io/Gen3DSR

comment: 3DV 2025 camera ready

♻ ☆ DH-Mamba: Exploring Dual-domain Hierarchical State Space Models for MRI Reconstruction

Yucong Meng, Zhiwei Yang, Zhijian Song, Yonghong Shi

The accelerated MRI reconstruction poses a challenging ill-posed inverse problem due to the significant undersampling in k-space. Deep neural networks, such as CNNs and ViTs, have shown substantial performance improvements for this task while encountering the dilemma between global receptive fields and efficient computation. To this end, this paper explores selective state space models (Mamba), a new paradigm for long-range dependency modeling with linear complexity, for efficient and effective MRI reconstruction. However, directly applying Mamba to MRI reconstruction faces three significant issues: (1) Mamba typically flattens 2D images into distinct 1D sequences along rows and columns, disrupting k-space's unique spectrum and leaving its potential in k-space learning unexplored. (2) Existing approaches adopt multi-directional lengthy scanning to unfold images at the pixel level, leading to long-range forgetting and high computational burden. (3) Mamba struggles with spatially-varying contents, resulting in limited diversity of local representations. To address these, we propose a dual-domain hierarchical Mamba for MRI reconstruction from the following perspectives: (1) We pioneer vision Mamba in k-space learning. A circular scanning is customized for spectrum unfolding, benefiting the global modeling of k-space. (2) We propose a hierarchical Mamba with an efficient scanning strategy in both image and k-space domains. It mitigates long-range forgetting and achieves a better trade-off between efficiency and performance. (3) We develop a local diversity enhancement module to improve the spatially-varying representation of Mamba. Extensive experiments are conducted on three public datasets for MRI reconstruction under various undersampling patterns. Comprehensive results demonstrate that our method significantly outperforms state-of-the-art methods with lower computational cost.

♻ ☆ MoMuSE: Momentum Multi-modal Target Speaker Extraction for Real-time Scenarios with Impaired Visual Cues

Junjie Li, Ke Zhang, Shuai Wang, Kong Aik Lee, Man-Wai Mak, Haizhou Li

Audio-visual Target Speaker Extraction (AV-TSE) aims to isolate the speech of a specific target speaker from an audio mixture using time-synchronized visual cues. In real-world scenarios, visual cues are not always available due to various impairments, which undermines the stability of AV-TSE. Despite this challenge, humans can maintain attentional momentum over time, even when the target speaker is not visible. In this paper, we introduce the Momentum Multi-modal target Speaker Extraction (MoMuSE), which retains a speaker identity momentum in memory, enabling the model to continuously track the target speaker. Designed for real-time inference, MoMuSE extracts the current speech window with guidance from both visual cues and dynamically updated speaker momentum. Experimental results demonstrate that MoMuSE exhibits significant improvement, particularly in scenarios with severe impairment of visual cues.

♻ ☆ Bayesian Learning-driven Prototypical Contrastive Loss for Class-Incremental Learning

Nisha L. Raichur, Lucas Heublein, Tobias Feigl, Alexander Rügamer, Christopher Mutschler, Felix Ott

The primary objective of methods in continual learning is to learn tasks in a sequential manner over time (sometimes from a stream of data), while mitigating the detrimental phenomenon of catastrophic forgetting. This paper proposes a method to learn an effective representation between previous and newly encountered class prototypes. We propose a prototypical network with a Bayesian learning-driven contrastive loss (BLCL), tailored specifically for class-incremental learning scenarios. We introduce a contrastive loss that incorporates novel classes into the latent representation by reducing intra-class and increasing inter-class distance. Our approach dynamically adapts the balance between the cross-entropy and contrastive loss functions with a Bayesian learning technique. Experimental results conducted on the CIFAR-10, CIFAR-100, and ImageNet100 datasets for image classification and images of a GNSS-based dataset for interference classification validate the efficacy of our method, showcasing its superiority over existing state-of-the-art approaches. Git: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/gnss_class_incremental_learning

comment: 27 pages, 22 figures

♻ ☆ Convolutional Kolmogorov-Arnold Networks

Alexander Dylan Bodner, Antonio Santiago Tepsich, Jack Natan Spolski, Santiago Pourteau

In this paper, we present Convolutional Kolmogorov-Arnold Networks, a novel architecture that integrates the learnable spline-based activation functions of Kolmogorov-Arnold Networks (KANs) into convolutional layers. By replacing traditional fixed-weight kernels with learnable non-linear functions, Convolutional KANs offer a significant improvement in parameter efficiency and expressive power over standard Convolutional Neural Networks (CNNs). We empirically evaluate Convolutional KANs on the Fashion-MNIST dataset, demonstrating competitive accuracy with up to 50% fewer parameters compared to baseline classic convolutions. This suggests that the KAN Convolution can effectively capture complex spatial relationships with fewer resources, offering a promising alternative for parameter-efficient deep learning models.

♻ ☆ LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning CVPR 2025

Xuan Liu, Xiaobin Chang

In continual learning (CL), catastrophic forgetting often arises due to feature drift. This challenge is particularly prominent in the exemplar-free continual learning (EFCL) setting, where samples from previous tasks cannot be retained, making it difficult to preserve prior knowledge. To address this issue, some EFCL methods aim to identify feature spaces that minimize the impact on previous tasks while accommodating new ones. However, they rely on static features or outdated statistics stored from old tasks, which prevents them from capturing the dynamic evolution of the feature space in CL, leading to performance degradation over time. In this paper, we introduce the Drift-Resistant Space (DRS), which effectively handles feature drifts without requiring explicit feature modeling or the storage of previous tasks. A novel parameter-efficient fine-tuning approach called Low-Rank Adaptation Subtraction (LoRA-) is proposed to develop the DRS. This method subtracts the LoRA weights of old tasks from the initial pre-trained weight before processing new task data to establish the DRS for model training. Therefore, LoRA- enhances stability, improves efficiency, and simplifies implementation. Furthermore, stabilizing feature drifts allows for better plasticity by learning with a triplet loss. Our method consistently achieves state-of-the-art results, especially for long task sequences, across multiple datasets.

comment: Accepted to CVPR 2025

♻ ☆ InPK: Infusing Prior Knowledge into Prompt for Vision-Language Models

Shuchang Zhou, Jiwei Wei, Shiyuan He, Yuyang Zhou, Chaoning Zhang, Jie Zou, Ning Xie, Yang Yang

Prompt tuning has become a popular strategy for adapting Vision-Language Models (VLMs) to zero/few-shot visual recognition tasks. Some prompting techniques introduce prior knowledge due to its richness, but when learnable tokens are randomly initialized and disconnected from prior knowledge, they tend to overfit on seen classes and struggle with domain shifts for unseen ones. To address this issue, we propose the InPK model, which infuses class-specific prior knowledge into the learnable tokens during initialization, thus enabling the model to explicitly focus on class-relevant information. Furthermore, to mitigate the weakening of class information by multi-layer encoders, we continuously reinforce the interaction between learnable tokens and prior knowledge across multiple feature levels. This progressive interaction allows the learnable tokens to better capture the fine-grained differences and universal visual concepts within prior knowledge, enabling the model to extract more discriminative and generalized text features. Even for unseen classes, the learned interaction allows the model to capture their common representations and infer their appropriate positions within the existing semantic structure. Moreover, we introduce a learnable text-to-vision projection layer to accommodate the text adjustments, ensuring better alignment of visual-text semantics. Extensive experiments on 11 recognition datasets show that InPK significantly outperforms state-of-the-art methods in multiple zero/few-shot image classification tasks.

♻ ☆ Cropper: Vision-Language Model for Image Cropping through In-Context Learning

Seung Hyun Lee, Jijun Jiang, Yiran Xu, Zhuofang Li, Junjie Ke, Yinxiao Li, Junfeng He, Steven Hickson, Katie Datsenko, Sangpil Kim, Ming-Hsuan Yang, Irfan Essa, Feng Yang

The goal of image cropping is to identify visually appealing crops in an image. Conventional methods are trained on specific datasets and fail to adapt to new requirements. Recent breakthroughs in large vision-language models (VLMs) enable visual in-context learning without explicit training. However, downstream tasks with VLMs remain under explored. In this paper, we propose an effective approach to leverage VLMs for image cropping. First, we propose an efficient prompt retrieval mechanism for image cropping to automate the selection of in-context examples. Second, we introduce an iterative refinement strategy to iteratively enhance the predicted crops. The proposed framework, we refer to as Cropper, is applicable to a wide range of cropping tasks, including free-form cropping, subject-aware cropping, and aspect ratio-aware cropping. Extensive experiments demonstrate that Cropper significantly outperforms state-of-the-art methods across several benchmarks.

♻ ☆ Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement AAAI 2025

Hyeonjin Kim, Jaejun Yoo

While pruning methods effectively maintain model performance without extra training costs, they often focus solely on preserving crucial connections, overlooking the impact of pruned weights on subsequent fine-tuning or distillation, leading to inefficiencies. Moreover, most compression techniques for generative models have been developed primarily for GANs, tailored to specific architectures like StyleGAN, and research into compressing Diffusion models has just begun. Even more, these methods are often applicable only to GANs or Diffusion models, highlighting the need for approaches that work across both model types. In this paper, we introduce Singular Value Scaling (SVS), a versatile technique for refining pruned weights, applicable to both model types. Our analysis reveals that pruned weights often exhibit dominant singular vectors, hindering fine-tuning efficiency and leading to suboptimal performance compared to random initialization. Our method enhances weight initialization by minimizing the disparities between singular values of pruned weights, thereby improving the fine-tuning process. This approach not only guides the compressed model toward superior solutions but also significantly speeds up fine-tuning. Extensive experiments on StyleGAN2, StyleGAN3 and DDPM demonstrate that SVS improves compression performance across model types without additional training costs. Our code is available at: https://github.com/LAIT-CVLab/Singular-Value-Scaling.

comment: Accepted to AAAI 2025

♻ ☆ MagicDistillation: Weak-to-Strong Video Distillation for Large-Scale Few-Step Synthesis

Shitong Shao, Hongwei Yi, Hanzhong Guo, Tian Ye, Daquan Zhou, Michael Lingelbach, Zhiqiang Xu, Zeke Xie

Recently, open-source video diffusion models (VDMs), such as WanX, Magic141 and HunyuanVideo, have been scaled to over 10 billion parameters. These large-scale VDMs have demonstrated significant improvements over smaller-scale VDMs across multiple dimensions, including enhanced visual quality and more natural motion dynamics. However, these models face two major limitations: (1) High inference overhead: Large-scale VDMs require approximately 10 minutes to synthesize a 28-step video on a single H100 GPU. (2) Limited in portrait video synthesis: Models like WanX-I2V and HunyuanVideo-I2V often produce unnatural facial expressions and movements in portrait videos. To address these challenges, we propose MagicDistillation, a novel framework designed to reduce inference overhead while ensuring the generalization of VDMs for portrait video synthesis. Specifically, we primarily use sufficiently high-quality talking video to fine-tune Magic141, which is dedicated to portrait video synthesis. We then employ LoRA to effectively and efficiently fine-tune the fake DiT within the step distillation framework known as distribution matching distillation (DMD). Following this, we apply weak-to-strong (W2S) distribution matching and minimize the discrepancy between the fake data distribution and the ground truth distribution, thereby improving the visual fidelity and motion dynamics of the synthesized videos. Experimental results on portrait video synthesis demonstrate the effectiveness of MagicDistillation, as our method surpasses Euler, LCM, and DMD baselines in both FID/FVD metrics and VBench. Moreover, MagicDistillation, requiring only 4 steps, also outperforms WanX-I2V (14B) and HunyuanVideo-I2V (13B) on visualization and VBench. Our project page is https://magicdistillation.github.io/MagicDistillation/.

♻ ☆ Beyond Walking: A Large-Scale Image-Text Benchmark for Text-based Person Anomaly Search

Shuyu Yang, Yaxiong Wang, Li Zhu, Zhedong Zheng

Text-based person search aims to retrieve specific individuals across camera networks using natural language descriptions. However, current benchmarks often exhibit biases towards common actions like walking or standing, neglecting the critical need for identifying abnormal behaviors in real-world scenarios. To meet such demands, we propose a new task, text-based person anomaly search, locating pedestrians engaged in both routine or anomalous activities via text. To enable the training and evaluation of this new task, we construct a large-scale image-text Pedestrian Anomaly Behavior (PAB) benchmark, featuring a broad spectrum of actions, e.g., running, performing, playing soccer, and the corresponding anomalies, e.g., lying, being hit, and falling of the same identity. The training set of PAB comprises 1,013,605 synthesized image-text pairs of both normalities and anomalies, while the test set includes 1,978 real-world image-text pairs. To validate the potential of PAB, we introduce a cross-modal pose-aware framework, which integrates human pose patterns with identity-based hard negative pair sampling. Extensive experiments on the proposed benchmark show that synthetic training data facilitates the fine-grained behavior retrieval, and the proposed pose-aware method arrives at 84.93% recall@1 accuracy, surpassing other competitive methods. The dataset, model, and code are available at https://github.com/Shuyu-XJTU/CMP.

♻ ☆ Data-free Universal Adversarial Perturbation with Pseudo-semantic Prior CVPR 2025

Chanhui Lee, Yeonghwan Song, Jeany Son

Data-free Universal Adversarial Perturbation (UAP) is an image-agnostic adversarial attack that deceives deep neural networks using a single perturbation generated solely from random noise without relying on data priors. However, traditional data-free UAP methods often suffer from limited transferability due to the absence of semantic content in random noise. To address this issue, we propose a novel data-free universal attack method that recursively extracts pseudo-semantic priors directly from the UAPs during training to enrich the semantic content within the data-free UAP framework. Our approach effectively leverages latent semantic information within UAPs via region sampling, enabling successful input transformations-typically ineffective in traditional data-free UAP methods due to the lack of semantic cues-and significantly enhancing black-box transferability. Furthermore, we introduce a sample reweighting technique to mitigate potential imbalances from random sampling and transformations, emphasizing hard examples less affected by the UAPs. Comprehensive experiments on ImageNet show that our method achieves state-of-the-art performance in average fooling rate by a substantial margin, notably improves attack transferability across various CNN architectures compared to existing data-free UAP methods, and even surpasses data-dependent UAP methods. Code is available at: https://github.com/ChnanChan/PSP-UAP.

comment: CVPR 2025

♻ ☆ Mitigating Cache Noise in Test-Time Adaptation for Large Vision-Language Models ICME 2025

Haotian Zhai, Xinyu Chen, Can Zhang, Tianming Sha, Ruirui Li

Test-time adaptation (TTA) of visual language models has recently attracted significant attention as a solution to the performance degradation caused by distribution shifts in downstream tasks. However, existing cache-based TTA methods have certain limitations. They mainly rely on the accuracy of cached feature labels, and the presence of noisy pseudo-labels can cause these features to deviate from their true distribution. This makes cache retrieval methods based on similarity matching highly sensitive to outliers or extreme samples. Moreover, current methods lack effective mechanisms to model class distributions, which limits their ability to fully exploit the potential of cached information. To address these challenges, we introduce a comprehensive and reliable caching mechanism and propose a novel zero-shot TTA method called "Cache, Residual, Gaussian" (CRG). This method not only employs learnable residual parameters to better align positive and negative visual prototypes with text prototypes, thereby optimizing the quality of cached features, but also incorporates Gaussian Discriminant Analysis (GDA) to dynamically model intra-class feature distributions, further mitigating the impact of noisy features. Experimental results on 13 benchmarks demonstrate that CRG outperforms state-of-the-art TTA methods, showcasing exceptional robustness and adaptability.

comment: Accepted by ICME 2025 and ICLR 2025 Workshop on Foundation Models in the Wild

♻ ☆ Interpreting Low-level Vision Models with Causal Effect Maps

Jinfan Hu, Jinjin Gu, Shiyao Yu, Fanghua Yu, Zheyuan Li, Zhiyuan You, Chaochao Lu, Chao Dong

Deep neural networks have significantly improved the performance of low-level vision tasks but also increased the difficulty of interpretability. A deep understanding of deep models is beneficial for both network design and practical reliability. To take up this challenge, we introduce causality theory to interpret low-level vision models and propose a model-/task-agnostic method called Causal Effect Map (CEM). With CEM, we can visualize and quantify the input-output relationships on either positive or negative effects. After analyzing various low-level vision tasks with CEM, we have reached several interesting insights, such as: (1) Using more information of input images (e.g., larger receptive field) does NOT always yield positive outcomes. (2) Attempting to incorporate mechanisms with a global receptive field (e.g., channel attention) into image denoising may prove futile. (3) Integrating multiple tasks to train a general model could encourage the network to prioritize local information over global context. Based on the causal effect theory, the proposed diagnostic tool can refresh our common knowledge and bring a deeper understanding of low-level vision models. Codes are available at https://github.com/J-FHu/CEM.

♻ ☆ Gaussian Eigen Models for Human Heads CVPR25

Wojciech Zielonka, Timo Bolkart, Thabo Beeler, Justus Thies

Current personalized neural head avatars face a trade-off: lightweight models lack detail and realism, while high-quality, animatable avatars require significant computational resources, making them unsuitable for commodity devices. To address this gap, we introduce Gaussian Eigen Models (GEM), which provide high-quality, lightweight, and easily controllable head avatars. GEM utilizes 3D Gaussian primitives for representing the appearance combined with Gaussian splatting for rendering. Building on the success of mesh-based 3D morphable face models (3DMM), we define GEM as an ensemble of linear eigenbases for representing the head appearance of a specific subject. In particular, we construct linear bases to represent the position, scale, rotation, and opacity of the 3D Gaussians. This allows us to efficiently generate Gaussian primitives of a specific head shape by a linear combination of the basis vectors, only requiring a low-dimensional parameter vector that contains the respective coefficients. We propose to construct these linear bases (GEM) by distilling high-quality compute-intense CNN-based Gaussian avatar models that can generate expression-dependent appearance changes like wrinkles. These high-quality models are trained on multi-view videos of a subject and are distilled using a series of principal component analyses. Once we have obtained the bases that represent the animatable appearance space of a specific human, we learn a regressor that takes a single RGB image as input and predicts the low-dimensional parameter vector that corresponds to the shown facial expression. In a series of experiments, we compare GEM's self-reenactment and cross-person reenactment results to state-of-the-art 3D avatar methods, demonstrating GEM's higher visual quality and better generalization to new expressions.

comment: Accepted to CVPR25 Website: https://zielon.github.io/gem/

♻ ☆ Synthetic Prior for Few-Shot Drivable Head Avatar Inversion CVPR25

Wojciech Zielonka, Stephan J. Garbin, Alexandros Lattas, George Kopanas, Paulo Gotardo, Thabo Beeler, Justus Thies, Timo Bolkart

We present SynShot, a novel method for the few-shot inversion of a drivable head avatar based on a synthetic prior. We tackle three major challenges. First, training a controllable 3D generative network requires a large number of diverse sequences, for which pairs of images and high-quality tracked meshes are not always available. Second, the use of real data is strictly regulated (e.g., under the General Data Protection Regulation, which mandates frequent deletion of models and data to accommodate a situation when a participant's consent is withdrawn). Synthetic data, free from these constraints, is an appealing alternative. Third, state-of-the-art monocular avatar models struggle to generalize to new views and expressions, lacking a strong prior and often overfitting to a specific viewpoint distribution. Inspired by machine learning models trained solely on synthetic data, we propose a method that learns a prior model from a large dataset of synthetic heads with diverse identities, expressions, and viewpoints. With few input images, SynShot fine-tunes the pretrained synthetic prior to bridge the domain gap, modeling a photorealistic head avatar that generalizes to novel expressions and viewpoints. We model the head avatar using 3D Gaussian splatting and a convolutional encoder-decoder that outputs Gaussian parameters in UV texture space. To account for the different modeling complexities over parts of the head (e.g., skin vs hair), we embed the prior with explicit control for upsampling the number of per-part primitives. Compared to SOTA monocular and GAN-based methods, SynShot significantly improves novel view and expression synthesis.

comment: Accepted to CVPR25 Website: https://zielon.github.io/synshot/

♻ ☆ Adaptive Multi-step Refinement Network for Robust Point Cloud Registration

Zhi Chen, Yufan Ren, Tong Zhang, Zheng Dang, Wenbing Tao, Sabine Süsstrunk, Mathieu Salzmann

Point Cloud Registration (PCR) estimates the relative rigid transformation between two point clouds of the same scene. Despite significant progress with learning-based approaches, existing methods still face challenges when the overlapping region between the two point clouds is small. In this paper, we propose an adaptive multi-step refinement network that refines the registration quality at each step by leveraging the information from the preceding step. To achieve this, we introduce a training procedure and a refinement network. Firstly, to adapt the network to the current step, we utilize a generalized one-way attention mechanism, which prioritizes the last step's estimated overlapping region, and we condition the network on step indices. Secondly, instead of training the network to map either random transformations or a fixed pre-trained model's estimations to the ground truth, we train it on transformations with varying registration qualities, ranging from accurate to inaccurate, thereby enhancing the network's adaptiveness and robustness. Despite its conceptual simplicity, our method achieves state-of-the-art performance on both the 3DMatch/3DLoMatch and KITTI benchmarks. Notably, on 3DLoMatch, our method reaches 80.4% recall rate, with an absolute improvement of 1.2%.

comment: Accepted at TMLR'25

♻ ☆ Resilient Sensor Fusion under Adverse Sensor Failures via Multi-Modal Expert Fusion CVPR 2025

Konyul Park, Yecheol Kim, Daehun Kim, Jun Won Choi

Modern autonomous driving perception systems utilize complementary multi-modal sensors, such as LiDAR and cameras. Although sensor fusion architectures enhance performance in challenging environments, they still suffer significant performance drops under severe sensor failures, such as LiDAR beam reduction, LiDAR drop, limited field of view, camera drop, and occlusion. This limitation stems from inter-modality dependencies in current sensor fusion frameworks. In this study, we introduce an efficient and robust LiDAR-camera 3D object detector, referred to as MoME, which can achieve robust performance through a mixture of experts approach. Our MoME fully decouples modality dependencies using three parallel expert decoders, which use camera features, LiDAR features, or a combination of both to decode object queries, respectively. We propose Multi-Expert Decoding (MED) framework, where each query is decoded selectively using one of three expert decoders. MoME utilizes an Adaptive Query Router (AQR) to select the most appropriate expert decoder for each query based on the quality of camera and LiDAR features. This ensures that each query is processed by the best-suited expert, resulting in robust performance across diverse sensor failure scenarios. We evaluated the performance of MoME on the nuScenes-R benchmark. Our MoME achieved state-of-the-art performance in extreme weather and sensor failure conditions, significantly outperforming the existing models across various sensor failure scenarios.

comment: Accepted to CVPR 2025

♻ ☆ RingMo-Aerial: An Aerial Remote Sensing Foundation Model With A Affine Transformation Contrastive Learning

Wenhui Diao, Haichen Yu, Kaiyue Kang, Tong Ling, Di Liu, Yingchao Feng, Hanbo Bi, Libo Ren, Xuexue Li, Yongqiang Mao, Xian Sun

Aerial Remote Sensing (ARS) vision tasks pose significant challenges due to the unique characteristics of their viewing angles. Existing research has primarily focused on algorithms for specific tasks, which have limited applicability in a broad range of ARS vision applications. This paper proposes the RingMo-Aerial model, aiming to fill the gap in foundation model research in the field of ARS vision. By introducing the Frequency-Enhanced Multi-Head Self-Attention (FE-MSA) mechanism and an affine transformation-based contrastive learning pre-training method, the model's detection capability for small targets is enhanced and optimized for the tilted viewing angles characteristic of ARS. Furthermore, the ARS-Adapter, an efficient parameter fine-tuning method, is proposed to improve the model's adaptability and effectiveness in various ARS vision tasks. Experimental results demonstrate that RingMo-Aerial achieves SOTA performance on multiple downstream tasks. This indicates the practicality and efficacy of RingMo-Aerial in enhancing the performance of ARS vision tasks.

♻ ☆ Bootstrap Your Own Views: Masked Ego-Exo Modeling for Fine-grained View-invariant Video Representations CVPR 2025

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov.

comment: CVPR 2025 Camera-ready, 18 pages, 7 figures, 9 tables

♻ ☆ 3D-GSW: 3D Gaussian Splatting for Robust Watermarking

Youngdong Jang, Hyunje Park, Feng Yang, Heeju Ko, Euijin Choo, Sangpil Kim

As 3D Gaussian Splatting (3D-GS) gains significant attention and its commercial usage increases, the need for watermarking technologies to prevent unauthorized use of the 3D-GS models and rendered images has become increasingly important. In this paper, we introduce a robust watermarking method for 3D-GS that secures copyright of both the model and its rendered images. Our proposed method remains robust against distortions in rendered images and model attacks while maintaining high rendering quality. To achieve these objectives, we present Frequency-Guided Densification (FGD), which removes 3D Gaussians based on their contribution to rendering quality, enhancing real-time rendering and the robustness of the message. FGD utilizes Discrete Fourier Transform to split 3D Gaussians in high-frequency areas, improving rendering quality. Furthermore, we employ a gradient mask for 3D Gaussians and design a wavelet-subband loss to enhance rendering quality. Our experiments show that our method embeds the message in the rendered images invisibly and robustly against various attacks, including model distortion. Our method achieves superior performance in both rendering quality and watermark robustness while improving real-time rendering efficiency. Project page: https://kuai-lab.github.io/cvpr20253dgsw/

♻ ☆ TransXNet: Learning Both Global and Local Dynamics with a Dual Dynamic Token Mixer for Visual Recognition

Meng Lou, Shu Zhang, Hong-Yu Zhou, Sibei Yang, Chuan Wu, Yizhou Yu

Recent studies have integrated convolutions into transformers to introduce inductive bias and improve generalization performance. However, the static nature of conventional convolution prevents it from dynamically adapting to input variations, resulting in a representation discrepancy between convolution and self-attention as the latter computes attention maps dynamically. Furthermore, when stacking token mixers that consist of convolution and self-attention to form a deep network, the static nature of convolution hinders the fusion of features previously generated by self-attention into convolution kernels. These two limitations result in a sub-optimal representation capacity of the entire network. To find a solution, we propose a lightweight Dual Dynamic Token Mixer (D-Mixer) to simultaneously learn global and local dynamics via computing input-dependent global and local aggregation weights. D-Mixer works by applying an efficient global attention module and an input-dependent depthwise convolution separately on evenly split feature segments, endowing the network with strong inductive bias and an enlarged receptive field. We use D-Mixer as the basic building block to design TransXNet, a novel hybrid CNN-Transformer vision backbone network that delivers compelling performance. In the ImageNet-1K classification, TransXNet-T surpasses Swin-T by 0.3% in top-1 accuracy while requiring less than half of the computational cost. Furthermore, TransXNet-S and TransXNet-B exhibit excellent model scalability, achieving top-1 accuracy of 83.8% and 84.6% respectively, with reasonable computational costs. Additionally, our proposed network architecture demonstrates strong generalization capabilities in various dense prediction tasks, outperforming other state-of-the-art networks while having lower computational costs. Code is publicly available at https://github.com/LMMMEng/TransXNet.

comment: Accepted by IEEE TNNLS

♻ ☆ Controllable Human Image Generation with Personalized Multi-Garments CVPR 2025

Yisol Choi, Sangkyung Kwak, Sihyun Yu, Hyungwon Choi, Jinwoo Shin

We present BootComp, a novel framework based on text-to-image diffusion models for controllable human image generation with multiple reference garments. Here, the main bottleneck is data acquisition for training: collecting a large-scale dataset of high-quality reference garment images per human subject is quite challenging, i.e., ideally, one needs to manually gather every single garment photograph worn by each human. To address this, we propose a data generation pipeline to construct a large synthetic dataset, consisting of human and multiple-garment pairs, by introducing a model to extract any reference garment images from each human image. To ensure data quality, we also propose a filtering strategy to remove undesirable generated data based on measuring perceptual similarities between the garment presented in human image and extracted garment. Finally, by utilizing the constructed synthetic dataset, we train a diffusion model having two parallel denoising paths that use multiple garment images as conditions to generate human images while preserving their fine-grained details. We further show the wide-applicability of our framework by adapting it to different types of reference-based generation in the fashion domain, including virtual try-on, and controllable human image generation with other conditions, e.g., pose, face, etc.

comment: CVPR 2025. Project page: https://omnious.github.io/BootComp

♻ ☆ HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation

Trong-Thuan Nguyen, Pha Nguyen, Jackson Cothren, Alper Yilmaz, Khoa Luu

Multimodal LLMs have advanced vision-language tasks but still struggle with understanding video scenes. To bridge this gap, Video Scene Graph Generation (VidSGG) has emerged to capture multi-object relationships across video frames. However, prior methods rely on pairwise connections, limiting their ability to handle complex multi-object interactions and reasoning. To this end, we propose Multimodal LLMs on a Scene HyperGraph (HyperGLM), promoting reasoning about multi-way interactions and higher-order relationships. Our approach uniquely integrates entity scene graphs, which capture spatial relationships between objects, with a procedural graph that models their causal transitions, forming a unified HyperGraph. Significantly, HyperGLM enables reasoning by injecting this unified HyperGraph into LLMs. Additionally, we introduce a new Video Scene Graph Reasoning (VSGR) dataset featuring 1.9M frames from third-person, egocentric, and drone views and supports five tasks: Scene Graph Generation, Scene Graph Anticipation, Video Question Answering, Video Captioning, and Relation Reasoning. Empirically, HyperGLM consistently outperforms state-of-the-art methods across five tasks, effectively modeling and reasoning complex relationships in diverse video scenes.

♻ ☆ RadSplat: Radiance Field-Informed Gaussian Splatting for Robust Real-Time Rendering with 900+ FPS 3DV

Michael Niemeyer, Fabian Manhardt, Marie-Julie Rakotosaona, Michael Oechsle, Daniel Duckworth, Rama Gosula, Keisuke Tateno, John Bates, Dominik Kaeser, Federico Tombari

Recent advances in view synthesis and real-time rendering have achieved photorealistic quality at impressive rendering speeds. While Radiance Field-based methods achieve state-of-the-art quality in challenging scenarios such as in-the-wild captures and large-scale scenes, they often suffer from excessively high compute requirements linked to volumetric rendering. Gaussian Splatting-based methods, on the other hand, rely on rasterization and naturally achieve real-time rendering but suffer from brittle optimization heuristics that underperform on more challenging scenes. In this work, we present RadSplat, a lightweight method for robust real-time rendering of complex scenes. Our main contributions are threefold. First, we use radiance fields as a prior and supervision signal for optimizing point-based scene representations, leading to improved quality and more robust optimization. Next, we develop a novel pruning technique reducing the overall point count while maintaining high quality, leading to smaller and more compact scene representations with faster inference speeds. Finally, we propose a novel test-time filtering approach that further accelerates rendering and allows to scale to larger, house-sized scenes. We find that our method enables state-of-the-art synthesis of complex captures at 900+ FPS.

comment: Project page at https://m-niemeyer.github.io/radsplat/ and presented at 3DV (Oral)

♻ ☆ Drag Your Gaussian: Effective Drag-Based Editing with Score Distillation for 3D Gaussian Splatting

Yansong Qu, Dian Chen, Xinyang Li, Xiaofan Li, Shengchuan Zhang, Liujuan Cao, Rongrong Ji

Recent advancements in 3D scene editing have been propelled by the rapid development of generative models. Existing methods typically utilize generative models to perform text-guided editing on 3D representations, such as 3D Gaussian Splatting (3DGS). However, these methods are often limited to texture modifications and fail when addressing geometric changes, such as editing a character's head to turn around. Moreover, such methods lack accurate control over the spatial position of editing results, as language struggles to precisely describe the extent of edits. To overcome these limitations, we introduce DYG, an effective 3D drag-based editing method for 3D Gaussian Splatting. It enables users to conveniently specify the desired editing region and the desired dragging direction through the input of 3D masks and pairs of control points, thereby enabling precise control over the extent of editing. DYG integrates the strengths of the implicit triplane representation to establish the geometric scaffold of the editing results, effectively overcoming suboptimal editing outcomes caused by the sparsity of 3DGS in the desired editing regions. Additionally, we incorporate a drag-based Latent Diffusion Model into our method through the proposed Drag-SDS loss function, enabling flexible, multi-view consistent, and fine-grained editing. Extensive experiments demonstrate that DYG conducts effective drag-based editing guided by control point prompts, surpassing other baselines in terms of editing effect and quality, both qualitatively and quantitatively. Visit our project page at https://quyans.github.io/Drag-Your-Gaussian.

comment: Visit our project page at https://quyans.github.io/Drag-Your-Gaussian

♻ ☆ DSU-Net:An Improved U-Net Model Based on DINOv2 and SAM2 with Multi-scale Cross-model Feature Enhancement

Yimin Xu, Fan Yang, Bin Xu

Despite the significant advancements in general image segmentation achieved by large-scale pre-trained foundation models (such as Meta's Segment Any-thing Model (SAM) series and DINOv2), their performance in specialized fields remains limited by two critical issues: the excessive training costs due to large model parameters, and the insufficient ability to represent specific domain characteristics. This paper proposes a multi-scale feature collabora-tion framework guided by DINOv2 for SAM2, with core innovations in three aspects: (1) Establishing a feature collaboration mechanism between DINOv2 and SAM2 backbones, where high-dimensional semantic features extracted by the self-supervised model guide multi-scale feature fusion; (2) Designing lightweight adapter modules and cross-modal, cross-layer feature fusion units to inject cross-domain knowledge while freezing the base model parameters; (3) Constructing a U-shaped network structure based on U-net, which utilizes attention mechanisms to achieve adaptive aggregation decoding of multi-granularity features. This framework surpasses existing state-of-the-art meth-ods in downstream tasks such as camouflage target detection and salient ob-ject detection, without requiring costly training processes. It provides a tech-nical pathway for efficient deployment of visual image segmentation, demon-strating significant application value in a wide range of downstream tasks and specialized fields within image segmentation.Project page: https://github.com/CheneyXuYiMin/SAM2DINO-Seg

♻ ☆ Mitigating Covariate Shift in Imitation Learning for Autonomous Vehicles Using Latent Space Generative World Models ICRA 2025

Alexander Popov, Alperen Degirmenci, David Wehr, Shashank Hegde, Ryan Oldja, Alexey Kamenev, Bertrand Douillard, David Nistér, Urs Muller, Ruchi Bhargava, Stan Birchfield, Nikolai Smolyanskiy

We propose the use of latent space generative world models to address the covariate shift problem in autonomous driving. A world model is a neural network capable of predicting an agent's next state given past states and actions. By leveraging a world model during training, the driving policy effectively mitigates covariate shift without requiring an excessive amount of training data. During end-to-end training, our policy learns how to recover from errors by aligning with states observed in human demonstrations, so that at runtime it can recover from perturbations outside the training distribution. Additionally, we introduce a novel transformer-based perception encoder that employs multi-view cross-attention and a learned scene query. We present qualitative and quantitative results, demonstrating significant improvements upon prior state of the art in closed-loop testing in the CARLA simulator, as well as showing the ability to handle perturbations in both CARLA and NVIDIA's DRIVE Sim.

comment: 8 pages, 6 figures, updated in March 2025, original published in September 2024, for ICRA 2025 submission, for associated video file, see https://youtu.be/7m3bXzlVQvU

♻ ☆ On-device Sora: Enabling Training-Free Diffusion-based Text-to-Video Generation for Mobile Devices

Bosung Kim, Kyuhwan Lee, Isu Jeong, Jungmin Cheon, Yeojin Lee, Seulki Lee

We present On-device Sora, the first model training-free solution for diffusion-based on-device text-to-video generation that operates efficiently on smartphone-grade devices. To address the challenges of diffusion-based text-to-video generation on computation- and memory-limited mobile devices, the proposed On-device Sora applies three novel techniques to pre-trained video generative models. First, Linear Proportional Leap (LPL) reduces the excessive denoising steps required in video diffusion through an efficient leap-based approach. Second, Temporal Dimension Token Merging (TDTM) minimizes intensive token-processing computation in attention layers by merging consecutive tokens along the temporal dimension. Third, Concurrent Inference with Dynamic Loading (CI-DL) dynamically partitions large models into smaller blocks and loads them into memory for concurrent model inference, effectively addressing the challenges of limited device memory. We implement On-device Sora on the iPhone 15 Pro, and the experimental evaluations show that it is capable of generating high-quality videos on the device, comparable to those produced by high-end GPUs. These results show that On-device Sora enables efficient and high-quality video generation on resource-constrained mobile devices. We envision the proposed On-device Sora as a significant first step toward democratizing state-of-the-art generative technologies, enabling video generation on commodity mobile and embedded devices without resource-intensive re-training for model optimization (compression). The code implementation is available at a GitHub repository(https://github.com/eai-lab/On-device-Sora).

♻ ☆ Know "No'' Better: A Data-Driven Approach for Enhancing Negation Awareness in CLIP

Junsung Park, Jungbeom Lee, Jongyoon Song, Sangwon Yu, Dahuin Jung, Sungroh Yoon

While CLIP has significantly advanced multimodal understanding by bridging vision and language, the inability to grasp negation - such as failing to differentiate concepts like "parking" from "no parking" - poses substantial challenges. By analyzing the data used in the public CLIP model's pre-training, we posit this limitation stems from a lack of negation-inclusive data. To address this, we introduce data generation pipelines that employ a large language model (LLM) and a multimodal LLM to produce negation-inclusive captions. Fine-tuning CLIP with data generated from our pipelines, we develop NegationCLIP, which enhances negation awareness while preserving the generality. Moreover, to enable a comprehensive evaluation of negation understanding, we propose NegRefCOCOg-a benchmark tailored to test VLMs' ability to interpret negation across diverse expressions and positions within a sentence. Experiments on various CLIP architectures validate the effectiveness of our data generation pipelines in enhancing CLIP's ability to perceive negation accurately. Additionally, NegationCLIP's enhanced negation awareness has practical applications across various multimodal tasks, demonstrated by performance gains in text-to-image generation and referring image segmentation.

♻ ☆ MultiBooth: Towards Generating All Your Concepts in an Image from Text AAAI 2025

Chenyang Zhu, Kai Li, Yue Ma, Chunming He, Xiu Li

This paper introduces MultiBooth, a novel and efficient technique for multi-concept customization in image generation from text. Despite the significant advancements in customized generation methods, particularly with the success of diffusion models, existing methods often struggle with multi-concept scenarios due to low concept fidelity and high inference cost. MultiBooth addresses these issues by dividing the multi-concept generation process into two phases: a single-concept learning phase and a multi-concept integration phase. During the single-concept learning phase, we employ a multi-modal image encoder and an efficient concept encoding technique to learn a concise and discriminative representation for each concept. In the multi-concept integration phase, we use bounding boxes to define the generation area for each concept within the cross-attention map. This method enables the creation of individual concepts within their specified regions, thereby facilitating the formation of multi-concept images. This strategy not only improves concept fidelity but also reduces additional inference cost. MultiBooth surpasses various baselines in both qualitative and quantitative evaluations, showcasing its superior performance and computational efficiency. Project Page: https://multibooth.github.io/

comment: To be published in AAAI 2025

♻ ☆ Vision-RWKV: Efficient and Scalable Visual Perception with RWKV-Like Architectures

Yuchen Duan, Weiyun Wang, Zhe Chen, Xizhou Zhu, Lewei Lu, Tong Lu, Yu Qiao, Hongsheng Li, Jifeng Dai, Wenhai Wang

Transformers have revolutionized computer vision and natural language processing, but their high computational complexity limits their application in high-resolution image processing and long-context analysis. This paper introduces Vision-RWKV (VRWKV), a model adapted from the RWKV model used in the NLP field with necessary modifications for vision tasks. Similar to the Vision Transformer (ViT), our model is designed to efficiently handle sparse inputs and demonstrate robust global processing capabilities, while also scaling up effectively, accommodating both large-scale parameters and extensive datasets. Its distinctive advantage lies in its reduced spatial aggregation complexity, which renders it exceptionally adept at processing high-resolution images seamlessly, eliminating the necessity for windowing operations. Our evaluations demonstrate that VRWKV surpasses ViT's performance in image classification and has significantly faster speeds and lower memory usage processing high-resolution inputs. In dense prediction tasks, it outperforms window-based models, maintaining comparable speeds. These results highlight VRWKV's potential as a more efficient alternative for visual perception tasks. Code is released at https://github.com/OpenGVLab/Vision-RWKV.

comment: Code is released at \url{https://github.com/OpenGVLab/Vision-RWKV}

♻ ☆ Enhancing Intent Understanding for Ambiguous prompt: A Human-Machine Co-Adaption Strategy

Yangfan He, Jianhui Wang, Yijin Wang, Kun Li, Yan Zhong, Xinyuan Song, Li Sun, Jingyuan Lu, Sida Li, Haoyuan Li, Jiayi Su, Jinhua Song, Miao Zhang, Tianyu Shi

Today's image generation systems are capable of producing realistic and high-quality images. However, user prompts often contain ambiguities, making it difficult for these systems to interpret users' actual intentions. Consequently, many users must modify their prompts several times to ensure the generated images meet their expectations. While some methods focus on enhancing prompts to make the generated images fit user needs, the model is still hard to understand users' real needs, especially for non-expert users. In this research, we aim to enhance the visual parameter-tuning process, making the model user-friendly for individuals without specialized knowledge and better understand user needs. We propose a human-machine co-adaption strategy using mutual information between the user's prompts and the pictures under modification as the optimizing target to make the system better adapt to user needs. We find that an improved model can reduce the necessity for multiple rounds of adjustments. We also collect multi-round dialogue datasets with prompts and images pairs and user intent. Various experiments demonstrate the effectiveness of the proposed method in our proposed dataset. Our annotation tools and several examples of our dataset are available at https://zenodo.org/records/14876029 for easier review. We will make open source our full dataset and code.

♻ ☆ Disentangled 4D Gaussian Splatting: Towards Faster and More Efficient Dynamic Scene Rendering

Hao Feng, Hao Sun, Wei Xie

Novel-view synthesis (NVS) for dynamic scenes from 2D images presents significant challenges due to the spatial complexity and temporal variability of such scenes. Recently, inspired by the remarkable success of NVS using 3D Gaussian Splatting (3DGS), researchers have sought to extend 3D Gaussian models to four dimensions (4D) for dynamic novel-view synthesis. However, methods based on 4D rotation and scaling introduce spatiotemporal deformation into the 4D covariance matrix, necessitating the slicing of 4D Gaussians into 3D Gaussians. This process increases redundant computations as timestamps change-an inherent characteristic of dynamic scene rendering. Additionally, performing calculations on a four-dimensional matrix is computationally intensive. In this paper, we introduce Disentangled 4D Gaussian Splatting (Disentangled4DGS), a novel representation and rendering approach that disentangles temporal and spatial deformations, thereby eliminating the reliance on 4D matrix computations. We extend the 3DGS rendering process to 4D, enabling the projection of temporal and spatial deformations into dynamic 2D Gaussians in ray space. Consequently, our method facilitates faster dynamic scene synthesis. Moreover, it reduces storage requirements by at least 4.5\% due to our efficient presentation method. Our approach achieves an unprecedented average rendering speed of 343 FPS at a resolution of $1352\times1014$ on an RTX 3090 GPU, with experiments across multiple benchmarks demonstrating its competitive performance in both monocular and multi-view scenarios.

♻ ☆ Boost Your Human Image Generation Model via Direct Preference Optimization CVPR

Sanghyeon Na, Yonggyu Kim, Hyunjoon Lee

Human image generation is a key focus in image synthesis due to its broad applications, but even slight inaccuracies in anatomy, pose, or details can compromise realism. To address these challenges, we explore Direct Preference Optimization (DPO), which trains models to generate preferred (winning) images while diverging from non-preferred (losing) ones. However, conventional DPO methods use generated images as winning images, limiting realism. To overcome this limitation, we propose an enhanced DPO approach that incorporates high-quality real images as winning images, encouraging outputs to resemble real images rather than generated ones. However, implementing this concept is not a trivial task. Therefore, our approach, HG-DPO (Human image Generation through DPO), employs a novel curriculum learning framework that gradually improves the output of the model toward greater realism, making training more feasible. Furthermore, HG-DPO effectively adapts to personalized text-to-image tasks, generating high-quality and identity-specific images, which highlights the practical value of our approach.

comment: CVPR`2025

♻ ☆ Context-Aware Weakly Supervised Image Manipulation Localization with SAM Refinement

Xinghao Wang, Tao Gong, Qi Chu, Bin Liu, Nenghai Yu

Malicious image manipulation poses societal risks, increasing the importance of effective image manipulation detection methods. Recent approaches in image manipulation detection have largely been driven by fully supervised approaches, which require labor-intensive pixel-level annotations. Thus, it is essential to explore weakly supervised image manipulation localization methods that only require image-level binary labels for training. However, existing weakly supervised image manipulation methods overlook the importance of edge information for accurate localization, leading to suboptimal localization performance. To address this, we propose a Context-Aware Boundary Localization (CABL) module to aggregate boundary features and learn context-inconsistency for localizing manipulated areas. Furthermore, by leveraging Class Activation Mapping (CAM) and Segment Anything Model (SAM), we introduce the CAM-Guided SAM Refinement (CGSR) module to generate more accurate manipulation localization maps. By integrating two modules, we present a novel weakly supervised framework based on a dual-branch Transformer-CNN architecture. Our method achieves outstanding localization performance across multiple datasets.

♻ ☆ Emphasizing Discriminative Features for Dataset Distillation in Complex Scenarios

Kai Wang, Zekai Li, Zhi-Qi Cheng, Samir Khaki, Ahmad Sajedi, Ramakrishna Vedantam, Konstantinos N Plataniotis, Alexander Hauptmann, Yang You

Dataset distillation has demonstrated strong performance on simple datasets like CIFAR, MNIST, and TinyImageNet but struggles to achieve similar results in more complex scenarios. In this paper, we propose EDF (emphasizes the discriminative features), a dataset distillation method that enhances key discriminative regions in synthetic images using Grad-CAM activation maps. Our approach is inspired by a key observation: in simple datasets, high-activation areas typically occupy most of the image, whereas in complex scenarios, the size of these areas is much smaller. Unlike previous methods that treat all pixels equally when synthesizing images, EDF uses Grad-CAM activation maps to enhance high-activation areas. From a supervision perspective, we downplay supervision signals that have lower losses, as they contain common patterns. Additionally, to help the DD community better explore complex scenarios, we build the Complex Dataset Distillation (Comp-DD) benchmark by meticulously selecting sixteen subsets, eight easy and eight hard, from ImageNet-1K. In particular, EDF consistently outperforms SOTA results in complex scenarios, such as ImageNet-1K subsets. Hopefully, more researchers will be inspired and encouraged to improve the practicality and efficacy of DD. Our code and benchmark will be made public at https://github.com/NUS-HPC-AI-Lab/EDF.

comment: 24 pages, 13 figures

♻ ☆ YOLO11 and Vision Transformers based 3D Pose Estimation of Immature Green Fruits in Commercial Apple Orchards for Robotic Thinning

Ranjan Sapkota, Manoj Karkee

In this study, a robust method for 3D pose estimation of immature green apples (fruitlets) in commercial orchards was developed, utilizing the YOLO11(or YOLOv11) object detection and pose estimation algorithm alongside Vision Transformers (ViT) for depth estimation (Dense Prediction Transformer (DPT) and Depth Anything V2). For object detection and pose estimation, performance comparisons of YOLO11 (YOLO11n, YOLO11s, YOLO11m, YOLO11l and YOLO11x) and YOLOv8 (YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l and YOLOv8x) were made under identical hyperparameter settings among the all configurations. It was observed that YOLO11n surpassed all configurations of YOLO11 and YOLOv8 in terms of box precision and pose precision, achieving scores of 0.91 and 0.915, respectively. Conversely, YOLOv8n exhibited the highest box and pose recall scores of 0.905 and 0.925, respectively. Regarding the mean average precision at 50\% intersection over union (mAP@50), YOLO11s led all configurations with a box mAP@50 score of 0.94, while YOLOv8n achieved the highest pose mAP@50 score of 0.96. In terms of image processing speed, YOLO11n outperformed all configurations with an impressive inference speed of 2.7 ms, significantly faster than the quickest YOLOv8 configuration, YOLOv8n, which processed images in 7.8 ms. Subsequent integration of ViTs for the green fruit's pose depth estimation revealed that Depth Anything V2 outperformed Dense Prediction Transformer in 3D pose length validation, achieving the lowest Root Mean Square Error (RMSE) of 1.52 and Mean Absolute Error (MAE) of 1.28, demonstrating exceptional precision in estimating immature green fruit lengths. Integration of YOLO11 and Depth Anything Model provides a promising solution to 3D pose estimation of immature green fruits for robotic thinning applications. (YOLOv11 pose detection, YOLOv11 Pose, YOLOv11 Keypoints detection, YOLOv11 pose estimation)

comment: 24 Pages, 13 Figures, 1 Table

♻ ☆ Hi3DGen: High-fidelity 3D Geometry Generation from Images via Normal Bridging

Chongjie Ye, Yushuang Wu, Ziteng Lu, Jiahao Chang, Xiaoyang Guo, Jiaqing Zhou, Hao Zhao, Xiaoguang Han

With the growing demand for high-fidelity 3D models from 2D images, existing methods still face significant challenges in accurately reproducing fine-grained geometric details due to limitations in domain gaps and inherent ambiguities in RGB images. To address these issues, we propose Hi3DGen, a novel framework for generating high-fidelity 3D geometry from images via normal bridging. Hi3DGen consists of three key components: (1) an image-to-normal estimator that decouples the low-high frequency image pattern with noise injection and dual-stream training to achieve generalizable, stable, and sharp estimation; (2) a normal-to-geometry learning approach that uses normal-regularized latent diffusion learning to enhance 3D geometry generation fidelity; and (3) a 3D data synthesis pipeline that constructs a high-quality dataset to support training. Extensive experiments demonstrate the effectiveness and superiority of our framework in generating rich geometric details, outperforming state-of-the-art methods in terms of fidelity. Our work provides a new direction for high-fidelity 3D geometry generation from images by leveraging normal maps as an intermediate representation.

comment: https://stable-x.github.io/Hi3DGen

♻ ☆ Enhancing Object Coherence in Layout-to-Image Synthesis

Yibin Wang, Changhai Zhou, Honghui Xu

Layout-to-image synthesis is an emerging technique in conditional image generation. It aims to generate complex scenes, where users require fine control over the layout of the objects in a scene. However, it remains challenging to control the object coherence, including semantic coherence (e.g., the cat looks at the flowers or not) and physical coherence (e.g., the hand and the racket should not be misaligned). In this paper, we propose a novel diffusion model with effective global semantic fusion (GSF) and self-similarity feature enhancement modules to guide the object coherence for this task. For semantic coherence, we argue that the image caption contains rich information for defining the semantic relationship within the objects in the images. Instead of simply employing cross-attention between captions and latent images, which addresses the highly relevant layout restriction and semantic coherence requirement separately and thus leads to unsatisfying results shown in our experiments, we develop GSF to fuse the supervision from the layout restriction and semantic coherence requirement and exploit it to guide the image synthesis process. Moreover, to improve the physical coherence, we develop a Self-similarity Coherence Attention (SCA) module to explicitly integrate local contextual physical coherence relation into each pixel's generation process. Specifically, we adopt a self-similarity map to encode the physical coherence restrictions and employ it to extract coherent features from text embedding. Through visualization of our self-similarity map, we explore the essence of SCA, revealing that its effectiveness is not only in capturing reliable physical coherence patterns but also in enhancing complex texture generation. Extensive experiments demonstrate the superiority of our proposed method.

comment: Code: https://github.com/CodeGoat24/EOCNet

♻ ☆ Exploring Cognitive Paradoxes in Video Games: A Quantum Mechanical Perspective

Ivan S. Maksymov, Ganna Pogrebna

This paper introduces a quantum-mechanical model that bridges the realms of cognition and quantum mechanics, offering a novel perspective on decision-making under risk and perceptual reversals. By integrating quantum theories addressing decision-theoretic anomalies with examples from immersive video games like "Deal or No Deal", we seek to elucidate complex human cognitive behaviours. Study 1 showcases the proposed quantum model's superiority over traditional decision-making approaches using the "Deal or No Deal" video game experiment. In Study 2, we apply our model to bistable perceptions, taking the Necker cube from the Necker game as a primary example. While previous works have hinted at connections between quantum mechanics and cognition, Study 3 provides a more tangible link, likening the physics that underpins quantum tunnelling to an eye blink's role in perceptual reversals. Conclusively, our model displays a promising ability to interpret diverse optical illusions and psychological phenomena, marking a significant stride in understanding human decision making.

♻ ☆ Head and Neck Tumor Segmentation of MRI from Pre- and Mid-radiotherapy with Pre-training, Data Augmentation and Dual Flow UNet

Litingyu Wang, Wenjun Liao, Shichuan Zhang, Guotai Wang

Head and neck tumors and metastatic lymph nodes are crucial for treatment planning and prognostic analysis. Accurate segmentation and quantitative analysis of these structures require pixel-level annotation, making automated segmentation techniques essential for the diagnosis and treatment of head and neck cancer. In this study, we investigated the effects of multiple strategies on the segmentation of pre-radiotherapy (pre-RT) and mid-radiotherapy (mid-RT) images. For the segmentation of pre-RT images, we utilized: 1) a fully supervised learning approach, and 2) the same approach enhanced with pre-trained weights and the MixUp data augmentation technique. For mid-RT images, we introduced a novel computational-friendly network architecture that features separate encoders for mid-RT images and registered pre-RT images with their labels. The mid-RT encoder branch integrates information from pre-RT images and labels progressively during the forward propagation. We selected the highest-performing model from each fold and used their predictions to create an ensemble average for inference. In the final test, our models achieved a segmentation performance of 82.38% for pre-RT and 72.53% for mid-RT on aggregated Dice Similarity Coefficient (DSC) as HiLab. Our code is available at https://github.com/WltyBY/HNTS-MRG2024_train_code.

♻ ☆ MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models ICLR 2025

Peng Xia, Siwei Han, Shi Qiu, Yiyang Zhou, Zhaoyang Wang, Wenhao Zheng, Zhaorun Chen, Chenhang Cui, Mingyu Ding, Linjie Li, Lijuan Wang, Huaxiu Yao

Interleaved multimodal comprehension and generation, enabling models to produce and interpret both images and text in arbitrary sequences, have become a pivotal area in multimodal learning. Despite significant advancements, the evaluation of this capability remains insufficient. Existing benchmarks suffer from limitations in data scale, scope, and evaluation depth, while current evaluation metrics are often costly or biased, lacking in reliability for practical applications. To address these challenges, we introduce MMIE, a large-scale knowledge-intensive benchmark for evaluating interleaved multimodal comprehension and generation in Large Vision-Language Models (LVLMs). MMIE comprises 20K meticulously curated multimodal queries, spanning 3 categories, 12 fields, and 102 subfields, including mathematics, coding, physics, literature, health, and arts. It supports both interleaved inputs and outputs, offering a mix of multiple-choice and open-ended question formats to evaluate diverse competencies. Moreover, we propose a reliable automated evaluation metric, leveraging a scoring model fine-tuned with human-annotated data and systematic evaluation criteria, aimed at reducing bias and improving evaluation accuracy. Extensive experiments demonstrate the effectiveness of our benchmark and metrics in providing a comprehensive evaluation of interleaved LVLMs. Specifically, we evaluate eight LVLMs, revealing that even the best models show significant room for improvement, with most achieving only moderate results. We believe MMIE will drive further advancements in the development of interleaved LVLMs. We publicly release our benchmark and code in https://mmie-bench.github.io/.

comment: ICLR 2025 Oral

♻ ☆ MovieBench: A Hierarchical Movie Level Dataset for Long Video Generation

Weijia Wu, Mingyu Liu, Zeyu Zhu, Xi Xia, Haoen Feng, Wen Wang, Kevin Qinghong Lin, Chunhua Shen, Mike Zheng Shou

Recent advancements in video generation models, like Stable Video Diffusion, show promising results, but primarily focus on short, single-scene videos. These models struggle with generating long videos that involve multiple scenes, coherent narratives, and consistent characters. Furthermore, there is no publicly available dataset tailored for the analysis, evaluation, and training of long video generation models. In this paper, we present MovieBench: A Hierarchical Movie-Level Dataset for Long Video Generation, which addresses these challenges by providing unique contributions: (1) movie-length videos featuring rich, coherent storylines and multi-scene narratives, (2) consistency of character appearance and audio across scenes, and (3) hierarchical data structure contains high-level movie information and detailed shot-level descriptions. Experiments demonstrate that MovieBench brings some new insights and challenges, such as maintaining character ID consistency across multiple scenes for various characters. The dataset will be public and continuously maintained, aiming to advance the field of long video generation. Data can be found at: https://weijiawu.github.io/MovieBench/.

comment: The project website is at: https://weijiawu.github.io/MovieBench/. Code: https://github.com/showlab/MovieBecnh

♻ ☆ Interpretable Few-shot Learning with Online Attribute Selection

Mohammad Reza Zarei, Majid Komeili

Few-shot learning (FSL) presents a challenging learning problem in which only a few samples are available for each class. Decision interpretation is more important in few-shot classification due to a greater chance of error compared to traditional classification. However, the majority of the previous FSL methods are black-box models. In this paper, we propose an inherently interpretable model for FSL based on human-friendly attributes. Previously, human-friendly attributes have been utilized to train models with the potential for human interaction and interpretability. However, such approaches are not directly extendible to the few-shot classification scenario. Moreover, we propose an online attribute selection mechanism to effectively filter out irrelevant attributes in each episode. The attribute selection mechanism improves accuracy and helps with interpretability by reducing the number of attributes that participate in each episode. We further propose a mechanism that automatically detects the episodes where the pool of available human-friendly attributes is insufficient, and subsequently augments it by engaging some learned unknown attributes. We demonstrate that the proposed method achieves results on par with black-box few-shot learning models on four widely used datasets. We also empirically evaluate the level of decision alignment between different models and human understanding and show that our model outperforms the comparison methods based on this criterion.

♻ ☆ Towards Geometric-Photometric Joint Alignment for Facial Mesh Registration

Xizhi Wang, Yaxiong Wang, Mengjian Li

This paper presents a Geometric-Photometric Joint Alignment~(GPJA) method, which aligns discrete human expressions at pixel-level accuracy by combining geometric and photometric information. Common practices for registering human heads typically involve aligning landmarks with facial template meshes using geometry processing approaches, but often overlook dense pixel-level photometric consistency. This oversight leads to inconsistent texture parametrization across different expressions, hindering the creation of topologically consistent head meshes widely used in movies and games. GPJA overcomes this limitation by leveraging differentiable rendering to align vertices with target expressions, achieving joint alignment in both geometry and photometric appearances automatically, without requiring semantic annotation or pre-aligned meshes for training. It features a holistic rendering alignment mechanism and a multiscale regularized optimization for robust convergence on large deformation. The method utilizes derivatives at vertex positions for supervision and employs a gradient-based algorithm which guarantees smoothness and avoids topological artifacts during the geometry evolution. Experimental results demonstrate faithful alignment under various expressions, surpassing the conventional non-rigid ICP-based methods and the state-of-the-art deep learning based method. In practical, our method generates meshes of the same subject across diverse expressions, all with the same texture parametrization. This consistency benefits face animation, re-parametrization, and other batch operations for face modeling and applications with enhanced efficiency.

♻ ☆ Skip-Vision: Efficient and Scalable Acceleration of Vision-Language Models via Adaptive Token Skipping

Weili Zeng, Ziyuan Huang, Kaixiang Ji, Yichao Yan

Transformer-based models have driven significant advancements in Multimodal Large Language Models (MLLMs), yet their computational costs surge drastically when scaling resolution, training data, and model parameters. A key bottleneck stems from the proliferation of visual tokens required for fine-grained image understanding. We propose Skip-Vision, a unified framework addressing both training and inference inefficiencies in vision-language models. On top of conventional token compression approaches, our method introduces two complementary acceleration strategies. For training acceleration, we observe that Feed-Forward Network (FFN) computations on visual tokens induce marginal feature updates. This motivates our Skip-FFN strategy, which bypasses FFN layers for redundant visual tokens. For inference acceleration, we design a selective KV-cache removal mechanism that prunes the skipped key-value pairs during decoding while preserving model performance. Experimental results demonstrate that Skip-Vision reduces training time by up to 35\%, inference FLOPs by 75\%, and latency by 45\%, while achieving comparable or superior performance to existing methods. Our work provides a practical solution for scaling high-performance MLLMs with enhanced efficiency.

♻ ☆ Will Pre-Training Ever End? A First Step Toward Next-Generation Foundation MLLMs via Self-Improving Systematic Cognition

Xiaoying Zhang, Da Peng, Yipeng Zhang, Zonghao Guo, Chengyue Wu, Chi Chen, Wei Ke, Helen Meng, Maosong Sun

Recent progress in (multimodal) large language models ((M)LLMs) has shifted focus from pre-training to inference-time compute scaling and post-training optimization, driven by concerns over limited high-quality real-world data. However, these strategies alone are insufficient for advancing model capabilities. We hypothesize that effective model improvement requires a strong synergy among pre-training, inference-time compute scaling, and post-training optimization. In this paper, we validate this hypothesis in the context of multimodal pre-training for foundation MLLM construction. We introduce Self-Improving cognition (SIcog), a self-learning framework for constructing next-generation foundation MLLMs by imparting multimodal knowledge and enhancing their systematic cognitive capabilities through multimodal pre-training with self-generated data. Specifically, we introduce Chain-of-Description, a step-by-step visual understanding method to improve comprehensive perception, and integrate structured chain-of-thought (CoT) reasoning to support in-depth multimodal reasoning. SIcog first equips a base model with systematic perception and reasoning using minimal external supervision. The enhanced model then generates candidate image captions and CoT-style reasoning responses for unlabeled images and image-question pairs across diverse tasks, which are curated through a self-consistency mechanism. These curated samples are subsequently used for large-scale multimodal pre-training, completing a self-learning cycle that strengthens the model's cognitive foundation. Extensive experiments demonstrate that SIcog produces next-generation foundation MLLMs with substantially improved multimodal cognition, outperforming prevailing pre-training approaches. These findings empirically establish SIcog as a promising framework for realizing a complete self-improving paradigm.

comment: 40 pages. Preprint, work in progress

♻ ☆ Diffusion-driven lensless fiber endomicroscopic quantitative phase imaging towards digital pathology

Zhaoqing Chen, Jiawei Sun, Xibin Yang, Xinyi Ye, Bin Zhao, Xuelong Li, Juergen Czarske

Lensless fiber endomicroscope is an emerging tool for in-vivo microscopic imaging, where quantitative phase imaging (QPI) can be utilized as a label-free method to enhance image contrast. However, existing single-shot phase reconstruction methods through lensless fiber endomicroscope typically perform well on simple images but struggle with complex microscopic structures. Here, we propose a speckle-conditioned diffusion model (SpecDiffusion), which reconstructs phase images directly from speckles captured at the detection side of a multi-core fiber (MCF). Unlike conventional neural networks, SpecDiffusion employs iterative phase denoising steps for speckle-driven phase reconstruction. The iteration scheme allows SpecDiffusion to break down the phase reconstruction process into multiple steps, gradually building up to the final phase image. This attribute alleviates the computation challenge at each step and enables the reconstruction of rich details in complex microscopic images. To validate its efficacy, we build an optical system to capture speckles from MCF and construct a dataset consisting of 100,000 paired images. SpecDiffusion provides high-fidelity phase reconstruction results and shows powerful generalization capacity for unseen objects, such as test charts and biological tissues, reducing the average mean absolute error of the reconstructed tissue images by 7 times. Furthermore, the reconstructed tissue images using SpecDiffusion shows higher accuracy in zero-shot cell segmentation tasks compared to the conventional method, demonstrating the potential for further cell morphology analysis through the learning-based lensless fiber endomicroscope. SpecDiffusion offers a precise and generalized method to phase reconstruction through scattering media, including MCFs, opening new perspective in lensless fiber endomicroscopic imaging.

♻ ☆ An interpretable approach to automating the assessment of biofouling in video footage

Evelyn J. Mannix, Bartholomew A. Woodham

Biofouling$\unicode{x2013}$communities of organisms that grow on hard surfaces immersed in water$\unicode{x2013}$provides a pathway for the spread of invasive marine species and diseases. To address this risk, international vessels are increasingly being obligated to provide evidence of their biofouling management practices. Verification that these activities are effective requires underwater inspections, using divers or underwater remotely operated vehicles (ROVs), and the collection and analysis of large amounts of imagery and footage. Automated assessment using computer vision techniques can significantly streamline this process, and this work shows how this challenge can be addressed efficiently and effectively using the interpretable Component Features (ComFe) approach with a DINOv2 Vision Transformer (ViT) foundation model. ComFe is able to obtain improved performance in comparison to previous non-interpretable Convolutional Neural Network (CNN) methods, with significantly fewer weights and greater transparency$\unicode{x2013}$through identifying which regions of the image contribute to the classification, and which images in the training data lead to that conclusion. All code, data and model weights are publicly released.

♻ ☆ DetailCLIP: Detail-Oriented CLIP for Fine-Grained Tasks ICLR 2025

Amin Karimi Monsefi, Kishore Prakash Sailaja, Ali Alilooee, Ser-Nam Lim, Rajiv Ramnath

In this paper, we introduce DetailCLIP: A Detail-Oriented CLIP to address the limitations of contrastive learning-based vision-language models, particularly CLIP, in handling detail-oriented and fine-grained tasks like segmentation. While CLIP and its variants excel in the global alignment of image and text representations, they often struggle to capture the fine-grained details necessary for precise segmentation. To overcome these challenges, we propose a novel framework that employs patch-level comparison of self-distillation and pixel-level reconstruction losses, enhanced with an attention-based token removal mechanism. This approach selectively retains semantically relevant tokens, enabling the model to focus on the image's critical regions aligned with the specific functions of our model, including textual information processing, patch comparison, and image reconstruction, ensuring that the model learns high-level semantics and detailed visual features. Our experiments demonstrate that DetailCLIP surpasses existing CLIP-based and traditional self-supervised learning (SSL) models in segmentation accuracy and exhibits superior generalization across diverse datasets. DetailCLIP represents a significant advancement in vision-language modeling, offering a robust solution for tasks that demand high-level semantic understanding and detailed feature extraction. https://github.com/KishoreP1/DetailCLIP.

comment: Accepted in SSI-FM Workshop of ICLR 2025

♻ ☆ Towards Adversarially Robust Dataset Distillation by Curvature Regularization

Eric Xue, Yijiang Li, Haoyang Liu, Peiran Wang, Yifan Shen, Haohan Wang

Dataset distillation (DD) allows datasets to be distilled to fractions of their original size while preserving the rich distributional information so that models trained on the distilled datasets can achieve a comparable accuracy while saving significant computational loads. Recent research in this area has been focusing on improving the accuracy of models trained on distilled datasets. In this paper, we aim to explore a new perspective of DD. We study how to embed adversarial robustness in distilled datasets, so that models trained on these datasets maintain the high accuracy and meanwhile acquire better adversarial robustness. We propose a new method that achieves this goal by incorporating curvature regularization into the distillation process with much less computational overhead than standard adversarial training. Extensive empirical experiments suggest that our method not only outperforms standard adversarial training on both accuracy and robustness with less computation overhead but is also capable of generating robust distilled datasets that can withstand various adversarial attacks.

comment: 14 pages, 3 figures

♻ ☆ GyralNet Subnetwork Partitioning via Differentiable Spectral Modularity Optimization

Yan Zhuang, Minheng Chen, Chao Cao, Tong Chen, Jing Zhang, Xiaowei Yu, Yanjun Lyu, Lu Zhang, Tianming Liu, Dajiang Zhu

Understanding the structural and functional organization of the human brain requires a detailed examination of cortical folding patterns, among which the three-hinge gyrus (3HG) has been identified as a key structural landmark. GyralNet, a network representation of cortical folding, models 3HGs as nodes and gyral crests as edges, highlighting their role as critical hubs in cortico-cortical connectivity. However, existing methods for analyzing 3HGs face significant challenges, including the sub-voxel scale of 3HGs at typical neuroimaging resolutions, the computational complexity of establishing cross-subject correspondences, and the oversimplification of treating 3HGs as independent nodes without considering their community-level relationships. To address these limitations, we propose a fully differentiable subnetwork partitioning framework that employs a spectral modularity maximization optimization strategy to modularize the organization of 3HGs within GyralNet. By incorporating topological structural similarity and DTI-derived connectivity patterns as attribute features, our approach provides a biologically meaningful representation of cortical organization. Extensive experiments on the Human Connectome Project (HCP) dataset demonstrate that our method effectively partitions GyralNet at the individual level while preserving the community-level consistency of 3HGs across subjects, offering a robust foundation for understanding brain connectivity.

comment: 10 pages, 3 figures

♻ ☆ VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding CVPR 2025

Chaoyu Li, Eun Woo Im, Pooyan Fazli

Multimodal large language models (MLLMs) have recently shown significant advancements in video understanding, excelling in content reasoning and instruction-following tasks. However, hallucination, where models generate inaccurate or misleading content, remains underexplored in the video domain. Building on the observation that MLLM visual encoders often fail to distinguish visually different yet semantically similar video pairs, we introduce VidHalluc, the largest benchmark designed to examine hallucinations in MLLMs for video understanding. It consists of 5,002 videos, paired to highlight cases prone to hallucinations. VidHalluc assesses hallucinations across three critical dimensions: (1) action, (2) temporal sequence, and (3) scene transition. Comprehensive testing shows that most MLLMs are vulnerable to hallucinations across these dimensions. Furthermore, we propose DINO-HEAL, a training-free method that reduces hallucinations by incorporating spatial saliency from DINOv2 to reweight visual features during inference. Our results show that DINO-HEAL consistently improves performance on VidHalluc, achieving an average improvement of 3.02% in mitigating hallucinations across all tasks. Both the VidHalluc benchmark and DINO-HEAL code are available at https://people-robots.github.io/vidhalluc.

comment: CVPR 2025

♻ ☆ Learning Color Equivariant Representations ICLR 2025

Yulong Yang, Felix O'Mahony, Christine Allen-Blanchette

In this paper, we introduce group convolutional neural networks (GCNNs) equivariant to color variation. GCNNs have been designed for a variety of geometric transformations from 2D and 3D rotation groups, to semi-groups such as scale. Despite the improved interpretability, accuracy and generalizability of these architectures, GCNNs have seen limited application in the context of perceptual quantities. Notably, the recent CEConv network uses a GCNN to achieve equivariance to hue transformations by convolving input images with a hue rotated RGB filter. However, this approach leads to invalid RGB values which break equivariance and degrade performance. We resolve these issues with a lifting layer that transforms the input image directly, thereby circumventing the issue of invalid RGB values and improving equivariance error by over three orders of magnitude. Moreover, we extend the notion of color equivariance to include equivariance to saturation and luminance shift. Our hue-, saturation-, luminance- and color-equivariant networks achieve strong generalization to out-of-distribution perceptual variations and improved sample efficiency over conventional architectures. We demonstrate the utility of our approach on synthetic and real world datasets where we consistently outperform competitive baselines.

comment: Accept to The 13th International Conference on Learning Representations (ICLR 2025)

♻ ☆ Interact with me: Joint Egocentric Forecasting of Intent to Interact, Attitude and Social Actions ICME

Tongfei Bian, Yiming Ma, Mathieu Chollet, Victor Sanchez, Tanaya Guha

For efficient human-agent interaction, an agent should proactively recognize their target user and prepare for upcoming interactions. We formulate this challenging problem as the novel task of jointly forecasting a person's intent to interact with the agent, their attitude towards the agent and the action they will perform, from the agent's (egocentric) perspective. So we propose \emph{SocialEgoNet} - a graph-based spatiotemporal framework that exploits task dependencies through a hierarchical multitask learning approach. SocialEgoNet uses whole-body skeletons (keypoints from face, hands and body) extracted from only 1 second of video input for high inference speed. For evaluation, we augment an existing egocentric human-agent interaction dataset with new class labels and bounding box annotations. Extensive experiments on this augmented dataset, named JPL-Social, demonstrate \emph{real-time} inference and superior performance (average accuracy across all tasks: 83.15\%) of our model outperforming several competitive baselines. The additional annotations and code will be available upon acceptance.

comment: Accepted at ICME, 2025

♻ ☆ Self-Calibrating 4D Novel View Synthesis from Monocular Videos Using Gaussian Splatting SC-4

Fang Li, Hao Zhang, Narendra Ahuja

Gaussian Splatting (GS) has significantly elevated scene reconstruction efficiency and novel view synthesis (NVS) accuracy compared to Neural Radiance Fields (NeRF), particularly for dynamic scenes. However, current 4D NVS methods, whether based on GS or NeRF, primarily rely on camera parameters provided by COLMAP and even utilize sparse point clouds generated by COLMAP for initialization, which lack accuracy as well are time-consuming. This sometimes results in poor dynamic scene representation, especially in scenes with large object movements, or extreme camera conditions e.g. small translations combined with large rotations. Some studies simultaneously optimize the estimation of camera parameters and scenes, supervised by additional information like depth, optical flow, etc. obtained from off-the-shelf models. Using this unverified information as ground truth can reduce robustness and accuracy, which does frequently occur for long monocular videos (with e.g. > hundreds of frames). We propose a novel approach that learns a high-fidelity 4D GS scene representation with self-calibration of camera parameters. It includes the extraction of 2D point features that robustly represent 3D structure, and their use for subsequent joint optimization of camera parameters and 3D structure towards overall 4D scene optimization. We demonstrate the accuracy and time efficiency of our method through extensive quantitative and qualitative experimental results on several standard benchmarks. The results show significant improvements over state-of-the-art methods for 4D novel view synthesis. The source code will be released soon at https://github.com/fangli333/SC-4DGS.

comment: GitHub Page: https://github.com/fangli333/SC-4DGS

♻ ☆ VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models CVPR 2025

Muchao Ye, Weiyang Liu, Pan He

The rapid advancement of vision-language models (VLMs) has established a new paradigm in video anomaly detection (VAD): leveraging VLMs to simultaneously detect anomalies and provide comprehendible explanations for the decisions. Existing work in this direction often assumes the complex reasoning required for VAD exceeds the capabilities of pretrained VLMs. Consequently, these approaches either incorporate specialized reasoning modules during inference or rely on instruction tuning datasets through additional training to adapt VLMs for VAD. However, such strategies often incur substantial computational costs or data annotation overhead. To address these challenges in explainable VAD, we introduce a verbalized learning framework named VERA that enables VLMs to perform VAD without model parameter modifications. Specifically, VERA automatically decomposes the complex reasoning required for VAD into reflections on simpler, more focused guiding questions capturing distinct abnormal patterns. It treats these reflective questions as learnable parameters and optimizes them through data-driven verbal interactions between learner and optimizer VLMs, using coarsely labeled training data. During inference, VERA embeds the learned questions into model prompts to guide VLMs in generating segment-level anomaly scores, which are then refined into frame-level scores via the fusion of scene and temporal contexts. Experimental results on challenging benchmarks demonstrate that the learned questions of VERA are highly adaptable, significantly improving both detection performance and explainability of VLMs for VAD.

comment: Accepted in CVPR 2025

♻ ☆ HaSPeR: An Image Repository for Hand Shadow Puppet Recognition

Syed Rifat Raiyan, Zibran Zarif Amio, Sabbir Ahmed

Hand shadow puppetry, also known as shadowgraphy or ombromanie, is a form of theatrical art and storytelling where hand shadows are projected onto flat surfaces to create illusions of living creatures. The skilled performers create these silhouettes by hand positioning, finger movements, and dexterous gestures to resemble shadows of animals and objects. Due to the lack of practitioners and a seismic shift in people's entertainment standards, this art form is on the verge of extinction. To facilitate its preservation and proliferate it to a wider audience, we introduce ${\rm H{\small A}SP{\small E}R}$, a novel dataset consisting of 15,000 images of hand shadow puppets across 15 classes extracted from both professional and amateur hand shadow puppeteer clips. We provide a detailed statistical analysis of the dataset and employ a range of pretrained image classification models to establish baselines. Our findings show a substantial performance superiority of skip-connected convolutional models over attention-based transformer architectures. We also find that lightweight models, such as MobileNetV2, suited for mobile applications and embedded devices, perform comparatively well. We surmise that such low-latency architectures can be useful in developing ombromanie teaching tools, and we create a prototype application to explore this surmission. Keeping the best-performing model ResNet34 under the limelight, we conduct comprehensive feature-spatial, explainability, and error analyses to gain insights into its decision-making process. To the best of our knowledge, this is the first documented dataset and research endeavor to preserve this dying art for future generations, with computer vision approaches. Our code and data will be publicly available.

comment: Submitted to Image and Vision Computing, 15 pages, 110 figures, 2 tables

♻ ☆ PTDiffusion: Free Lunch for Generating Optical Illusion Hidden Pictures with Phase-Transferred Diffusion Model CVPR 2025

Xiang Gao, Shuai Yang, Jiaying Liu

Optical illusion hidden picture is an interesting visual perceptual phenomenon where an image is cleverly integrated into another picture in a way that is not immediately obvious to the viewer. Established on the off-the-shelf text-to-image (T2I) diffusion model, we propose a novel training-free text-guided image-to-image (I2I) translation framework dubbed as \textbf{P}hase-\textbf{T}ransferred \textbf{Diffusion} Model (PTDiffusion) for hidden art syntheses. PTDiffusion harmoniously embeds an input reference image into arbitrary scenes described by the text prompts, producing illusion images exhibiting hidden visual cues of the reference image. At the heart of our method is a plug-and-play phase transfer mechanism that dynamically and progressively transplants diffusion features' phase spectrum from the denoising process to reconstruct the reference image into the one to sample the generated illusion image, realizing deep fusion of the reference structural information and the textual semantic information in the diffusion model latent space. Furthermore, we propose asynchronous phase transfer to enable flexible control to the degree of hidden content discernability. Our method bypasses any model training and fine-tuning process, all while substantially outperforming related text-guided I2I methods in image generation quality, text fidelity, visual discernibility, and contextual naturalness for illusion picture synthesis, as demonstrated by extensive qualitative and quantitative experiments. Our project is publically available at \href{https://xianggao1102.github.io/PTDiffusion_webpage/}{this web page}.

comment: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025)

♻ ☆ MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine ICLR 2025

Yunfei Xie, Ce Zhou, Lang Gao, Juncheng Wu, Xianhang Li, Hong-Yu Zhou, Sheng Liu, Lei Xing, James Zou, Cihang Xie, Yuyin Zhou

This paper introduces MedTrinity-25M, a comprehensive, large-scale multimodal dataset for medicine, covering over 25 million images across 10 modalities with multigranular annotations for more than 65 diseases. These multigranular annotations encompass both global information, such as modality and organ detection, and local information like ROI analysis, lesion texture, and region-wise correlations. Unlike the existing multimodal datasets, which are limited by the availability of image-text pairs, we have developed the first automated pipeline that scales up multimodal data by generating multigranular visual and textual annotations in the form of image-ROI-description triplets without the need for any paired text descriptions. Specifically, data from over 30 different sources have been collected, preprocessed, and grounded using domain-specific expert models to identify ROIs related to abnormal regions. We then build a comprehensive knowledge base and prompt multimodal large language models to perform retrieval-augmented generation with the identified ROIs as guidance, resulting in multigranular textual descriptions. Compared to existing datasets, MedTrinity-25M provides the most enriched annotations, supporting a comprehensive range of multimodal tasks such as captioning and report generation, as well as vision-centric tasks like classification and segmentation. We propose LLaVA-Tri by pretraining LLaVA on MedTrinity-25M, achieving state-of-the-art performance on VQA-RAD, SLAKE, and PathVQA, surpassing representative SOTA multimodal large language models. Furthermore, MedTrinity-25M can also be utilized to support large-scale pre-training of multimodal medical AI models, contributing to the development of future foundation models in the medical domain. We will make our dataset available.

comment: The dataset is publicly available at https://yunfeixie233.github.io/MedTrinity-25M/. Accepted to ICLR 2025

♻ ☆ BIGbench: A Unified Benchmark for Evaluating Multi-dimensional Social Biases in Text-to-Image Models

Hanjun Luo, Haoyu Huang, Ziye Deng, Xinfeng Li, Hewei Wang, Yingbin Jin, Yang Liu, Wenyuan Xu, Zuozhu Liu

Text-to-Image (T2I) generative models are becoming increasingly crucial due to their ability to generate high-quality images, but also raise concerns about social biases, particularly in human image generation. Sociological research has established systematic classifications of bias. Yet, existing studies on bias in T2I models largely conflate different types of bias, impeding methodological progress. In this paper, we introduce BIGbench, a unified benchmark for Biases of Image Generation, featuring a carefully designed dataset. Unlike existing benchmarks, BIGbench classifies and evaluates biases across four dimensions to enable a more granular evaluation and deeper analysis. Furthermore, BIGbench applies advanced multi-modal large language models to achieve fully automated and highly accurate evaluations. We apply BIGbench to evaluate eight representative T2I models and three debiasing methods. Our human evaluation results by trained evaluators from different races underscore BIGbench's effectiveness in aligning images and identifying various biases. Moreover, our study also reveals new research directions about biases with insightful analysis of our results. Our work is openly accessible at https://github.com/BIGbench2024/BIGbench2024/.

comment: arXiv admin note: substantial text overlap with arXiv:2405.17814

☆ A Systematic Evaluation of LLM Strategies for Mental Health Text Analysis: Fine-tuning vs. Prompt Engineering vs. RAG

Arshia Kermani, Veronica Perez-Rosas, Vangelis Metsis

This study presents a systematic comparison of three approaches for the analysis of mental health text using large language models (LLMs): prompt engineering, retrieval augmented generation (RAG), and fine-tuning. Using LLaMA 3, we evaluate these approaches on emotion classification and mental health condition detection tasks across two datasets. Fine-tuning achieves the highest accuracy (91% for emotion classification, 80% for mental health conditions) but requires substantial computational resources and large training sets, while prompt engineering and RAG offer more flexible deployment with moderate performance (40-68% accuracy). Our findings provide practical insights for implementing LLM-based solutions in mental health applications, highlighting the trade-offs between accuracy, computational requirements, and deployment flexibility.

☆ Rec-R1: Bridging Generative Large Language Models and User-Centric Recommendation Systems via Reinforcement Learning

Jiacheng Lin, Tian Wang, Kun Qian

We propose Rec-R1, a general reinforcement learning framework that bridges large language models (LLMs) with recommendation systems through closed-loop optimization. Unlike prompting and supervised fine-tuning (SFT), Rec-R1 directly optimizes LLM generation using feedback from a fixed black-box recommendation model, without relying on synthetic SFT data from proprietary models such as GPT-4o. This avoids the substantial cost and effort required for data distillation. To verify the effectiveness of Rec-R1, we evaluate it on two representative tasks: product search and sequential recommendation. Experimental results demonstrate that Rec-R1 not only consistently outperforms prompting- and SFT-based methods, but also achieves significant gains over strong discriminative baselines, even when used with simple retrievers such as BM25. Moreover, Rec-R1 preserves the general-purpose capabilities of the LLM, unlike SFT, which often impairs instruction-following and reasoning. These findings suggest Rec-R1 as a promising foundation for continual task-specific adaptation without catastrophic forgetting.

☆ Combining Query Performance Predictors: A Reproducibility Study

Sourav Saha, Suchana Datta, Dwaipayan Roy, Mandar Mitra, Derek Greene

A large number of approaches to Query Performance Prediction (QPP) have been proposed over the last two decades. As early as 2009, Hauff et al. [28] explored whether different QPP methods may be combined to improve prediction quality. Since then, significant research has been done both on QPP approaches, as well as their evaluation. This study revisits Hauff et al.s work to assess the reproducibility of their findings in the light of new prediction methods, evaluation metrics, and datasets. We expand the scope of the earlier investigation by: (i) considering post-retrieval methods, including supervised neural techniques (only pre-retrieval techniques were studied in [28]); (ii) using sMARE for evaluation, in addition to the traditional correlation coefficients and RMSE; and (iii) experimenting with additional datasets (Clueweb09B and TREC DL). Our results largely support previous claims, but we also present several interesting findings. We interpret these findings by taking a more nuanced look at the correlation between QPP methods, examining whether they capture diverse information or rely on overlapping factors.

☆ Music Information Retrieval on Representative Mexican Folk Vocal Melodies Through MIDI Feature Extraction

Mario Alberto Vallejo Reyes

This study analyzes representative Mexican folk vocal melodies using MIDI feature extraction, examining ambitus, pitch-class entropy, and interval distribution. It also explores the relationship between these features and song popularity, as measured by Spotify plays. The study employs MATLAB and the MIDI Toolbox for extracting musical features and performing statistical analysis. The findings reveal a significant variation in ambitus, with values ranging from 8 to 27 semitones, indicating a diverse compositional style and vocal demand across the genre. The analysis of pitch-class entropy showcases a broad spectrum of melodic complexity, with Armando Manzanero's `Somos Novios' displaying the highest entropy, suggesting varied and complex melodic structures, while traditional pieces like `La Bamba' exhibit lower entropy, indicating simpler, more repetitive patterns. The interval distribution predominantly features prime intervals (P1), major and minor seconds (M2, m2), pointing to a compositional preference for close, contiguous intervals that contribute to the melodies' accessibility and appeal. Statistical analysis do not establish a significant correlation between the ambitus or entropy and the number of Spotify plays.

comment: 10 pages, 5 figures, 2 tables

☆ Text2Tracks: Prompt-based Music Recommendation via Generative Retrieval

Enrico Palumbo, Gustavo Penha, Andreas Damianou, José Luis Redondo García, Timothy Christopher Heath, Alice Wang, Hugues Bouchard, Mounia Lalmas

In recent years, Large Language Models (LLMs) have enabled users to provide highly specific music recommendation requests using natural language prompts (e.g. "Can you recommend some old classics for slow dancing?"). In this setup, the recommended tracks are predicted by the LLM in an autoregressive way, i.e. the LLM generates the track titles one token at a time. While intuitive, this approach has several limitation. First, it is based on a general purpose tokenization that is optimized for words rather than for track titles. Second, it necessitates an additional entity resolution layer that matches the track title to the actual track identifier. Third, the number of decoding steps scales linearly with the length of the track title, slowing down inference. In this paper, we propose to address the task of prompt-based music recommendation as a generative retrieval task. Within this setting, we introduce novel, effective, and efficient representations of track identifiers that significantly outperform commonly used strategies. We introduce Text2Tracks, a generative retrieval model that learns a mapping from a user's music recommendation prompt to the relevant track IDs directly. Through an offline evaluation on a dataset of playlists with language inputs, we find that (1) the strategy to create IDs for music tracks is the most important factor for the effectiveness of Text2Tracks and semantic IDs significantly outperform commonly used strategies that rely on song titles as identifiers (2) provided with the right choice of track identifiers, Text2Tracks outperforms sparse and dense retrieval solutions trained to retrieve tracks from language prompts.

☆ On the Reproducibility of Learned Sparse Retrieval Adaptations for Long Documents ECIR 2025

Emmanouil Georgios Lionis, Jia-Huei Ju

Document retrieval is one of the most challenging tasks in Information Retrieval. It requires handling longer contexts, often resulting in higher query latency and increased computational overhead. Recently, Learned Sparse Retrieval (LSR) has emerged as a promising approach to address these challenges. Some have proposed adapting the LSR approach to longer documents by aggregating segmented document using different post-hoc methods, including n-grams and proximity scores, adjusting representations, and learning to ensemble all signals. In this study, we aim to reproduce and examine the mechanisms of adapting LSR for long documents. Our reproducibility experiments confirmed the importance of specific segments, with the first segment consistently dominating document retrieval performance. Furthermore, We re-evaluate recently proposed methods -- ExactSDM and SoftSDM -- across varying document lengths, from short (up to 2 segments) to longer (3+ segments). We also designed multiple analyses to probe the reproduced methods and shed light on the impact of global information on adapting LSR to longer contexts. The complete code and implementation for this project is available at: https://github.com/lionisakis/Reproducibilitiy-lsr-long.

comment: This is a preprint of our paper accepted at ECIR 2025

☆ Get the Agents Drunk: Memory Perturbations in Autonomous Agent-based Recommender Systems

Shiyi Yang, Zhibo Hu, Chen Wang, Tong Yu, Xiwei Xu, Liming Zhu, Lina Yao

Large language model-based agents are increasingly used in recommender systems (Agent4RSs) to achieve personalized behavior modeling. Specifically, Agent4RSs introduces memory mechanisms that enable the agents to autonomously learn and self-evolve from real-world interactions. However, to the best of our knowledge, how robust Agent4RSs are remains unexplored. As such, in this paper, we propose the first work to attack Agent4RSs by perturbing agents' memories, not only to uncover their limitations but also to enhance their security and robustness, ensuring the development of safer and more reliable AI agents. Given the security and privacy concerns, it is more practical to launch attacks under a black-box setting, where the accurate knowledge of the victim models cannot be easily obtained. Moreover, the practical attacks are often stealthy to maximize the impact. To this end, we propose a novel practical attack framework named DrunkAgent. DrunkAgent consists of a generation module, a strategy module, and a surrogate module. The generation module aims to produce effective and coherent adversarial textual triggers, which can be used to achieve attack objectives such as promoting the target items. The strategy module is designed to `get the target agents drunk' so that their memories cannot be effectively updated during the interaction process. As such, the triggers can play the best role. Both of the modules are optimized on the surrogate module to improve the transferability and imperceptibility of the attacks. By identifying and analyzing the vulnerabilities, our work provides critical insights that pave the way for building safer and more resilient Agent4RSs. Extensive experiments across various real-world datasets demonstrate the effectiveness of DrunkAgent.

☆ Finding Interest Needle in Popularity Haystack: Improving Retrieval by Modeling Item Exposure

Amit Jaspal, Rahul Agarwal

Recommender systems operate in closed feedback loops, where user interactions reinforce popularity bias, leading to over-recommendation of already popular items while under-exposing niche or novel content. Existing bias mitigation methods, such as Inverse Propensity Scoring (IPS) and Off- Policy Correction (OPC), primarily operate at the ranking stage or during training, lacking explicit real-time control over exposure dynamics. In this work, we introduce an exposure- aware retrieval scoring approach, which explicitly models item exposure probability and adjusts retrieval-stage ranking at inference time. Unlike prior work, this method decouples exposure effects from engagement likelihood, enabling controlled trade-offs between fairness and engagement in large-scale recommendation platforms. We validate our approach through online A/B experiments in a real-world video recommendation system, demonstrating a 25% increase in uniquely retrieved items and a 40% reduction in the dominance of over-popular content, all while maintaining overall user engagement levels. Our results establish a scalable, deployable solution for mitigating popularity bias at the retrieval stage, offering a new paradigm for bias-aware personalization.

comment: 2 pages

♻ ☆ Comparison of Metadata Representation Models for Knowledge Graph Embeddings

Shusaku Egami, Kyoumoto Matsushita, Takanori Ugai, Ken Fukuda

Hyper-relational Knowledge Graphs (HRKGs) extend traditional KGs beyond binary relations, enabling the representation of contextual, provenance, and temporal information in domains, such as historical events, sensor data, video content, and narratives. HRKGs can be structured using several Metadata Representation Models (MRMs), including Reification (REF), Singleton Property (SGP), and RDF-star (RDR). However, the effects of different MRMs on KG Embedding (KGE) and Link Prediction (LP) models remain unclear. This study evaluates MRMs in the context of LP tasks, identifies the limitations of existing evaluation frameworks, and introduces a new task that ensures fair comparisons across MRMs. Furthermore, we propose a framework that effectively reflects the knowledge representations of the three MRMs in latent space. Experiments on two types of datasets reveal that REF performs well in simple HRKGs, whereas SGP is less effective. However, in complex HRKGs, the differences among MRMs in the LP tasks are minimal. Our findings contribute to an optimal knowledge representation strategy for HRKGs in LP tasks.

comment: 11 pages, 9 Figures

♻ ☆ Can Zero-Shot Commercial APIs Deliver Regulatory-Grade Clinical Text DeIdentification? ECIR 2025

Veysel Kocaman, Muhammed Santas, Yigit Gul, Mehmet Butgul, David Talby

We evaluate the performance of four leading solutions for de-identification of unstructured medical text - Azure Health Data Services, AWS Comprehend Medical, OpenAI GPT-4o, and John Snow Labs - on a ground truth dataset of 48 clinical documents annotated by medical experts. The analysis, conducted at both entity-level and token-level, suggests that John Snow Labs' Medical Language Models solution achieves the highest accuracy, with a 96% F1-score in protected health information (PHI) detection, outperforming Azure (91%), AWS (83%), and GPT-4o (79%). John Snow Labs is not only the only solution which achieves regulatory-grade accuracy (surpassing that of human experts) but is also the most cost-effective solution: It is over 80% cheaper compared to Azure and GPT-4o, and is the only solution not priced by token. Its fixed-cost local deployment model avoids the escalating per-request fees of cloud-based services, making it a scalable and economical choice.

comment: 14 pages, accepted at Text2Story Workshop at ECIR 2025

☆ RIG: Synergizing Reasoning and Imagination in End-to-End Generalist Policy

Zhonghan Zhao, Wenwei Zhang, Haian Huang, Kuikun Liu, Jianfei Gao, Gaoang Wang, Kai Chen

Reasoning before action and imagining potential outcomes (i.e., world models) are essential for embodied agents operating in complex open-world environments. Yet, prior work either incorporates only one of these abilities in an end-to-end agent or integrates multiple specialized models into an agent system, limiting the learning efficiency and generalization of the policy. Thus, this paper makes the first attempt to synergize Reasoning and Imagination in an end-to-end Generalist policy, termed RIG. To train RIG in an end-to-end manner, we construct a data pipeline that progressively integrates and enriches the content of imagination and reasoning in the trajectories collected from existing agents. The joint learning of reasoning and next image generation explicitly models the inherent correlation between reasoning, action, and dynamics of environments, and thus exhibits more than $17\times$ sample efficiency improvements and generalization in comparison with previous works. During inference, RIG first reasons about the next action, produces potential action, and then predicts the action outcomes, which offers the agent a chance to review and self-correct based on the imagination before taking real actions. Experimental results show that the synergy of reasoning and imagination not only improves the robustness, generalization, and interoperability of generalist policy but also enables test-time scaling to enhance overall performance.

☆ UniOcc: A Unified Benchmark for Occupancy Forecasting and Prediction in Autonomous Driving

Yuping Wang, Xiangyu Huang, Xiaokang Sun, Mingxuan Yan, Shuo Xing, Zhengzhong Tu, Jiachen Li

We introduce UniOcc, a comprehensive, unified benchmark for occupancy forecasting (i.e., predicting future occupancies based on historical information) and current-frame occupancy prediction from camera images. UniOcc unifies data from multiple real-world datasets (i.e., nuScenes, Waymo) and high-fidelity driving simulators (i.e., CARLA, OpenCOOD), which provides 2D/3D occupancy labels with per-voxel flow annotations and support for cooperative autonomous driving. In terms of evaluation, unlike existing studies that rely on suboptimal pseudo labels for evaluation, UniOcc incorporates novel metrics that do not depend on ground-truth occupancy, enabling robust assessment of additional aspects of occupancy quality. Through extensive experiments on state-of-the-art models, we demonstrate that large-scale, diverse training data and explicit flow information significantly enhance occupancy prediction and forecasting performance.

comment: 14 pages; Dataset: https://huggingface.co/datasets/tasl-lab/uniocc; Code: https://github.com/tasl-lab/UniOcc

☆ Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1

Yi Chen, Yuying Ge, Rui Wang, Yixiao Ge, Lu Qiu, Ying Shan, Xihui Liu

Recent advancements in Chain of Thought (COT) generation have significantly improved the reasoning capabilities of Large Language Models (LLMs), with reinforcement learning (RL) emerging as an effective post-training approach. Multimodal Large Language Models (MLLMs) inherit this reasoning potential but remain underexplored in tasks requiring both perception and logical reasoning. To address this, we introduce SEED-Bench-R1, a benchmark designed to systematically evaluate post-training methods for MLLMs in video understanding. It includes intricate real-world videos and complex everyday planning tasks in the format of multiple-choice questions, requiring sophisticated perception and reasoning. SEED-Bench-R1 assesses generalization through a three-level hierarchy: in-distribution, cross-environment, and cross-environment-task scenarios, equipped with a large-scale training dataset with easily verifiable ground-truth answers. Using Qwen2-VL-Instruct-7B as a base model, we compare RL with supervised fine-tuning (SFT), demonstrating RL's data efficiency and superior performance on both in-distribution and out-of-distribution tasks, even outperforming SFT on general video understanding benchmarks like LongVideoBench. Our detailed analysis reveals that RL enhances visual perception but often produces less logically coherent reasoning chains. We identify key limitations such as inconsistent reasoning and overlooked visual cues, and suggest future improvements in base model reasoning, reward modeling, and RL robustness against noisy signals.

comment: Technical Report (In Progress); Code released at: https://github.com/TencentARC/SEED-Bench-R1

☆ Policy Gradient for LQR with Domain Randomization

Tesshu Fujinami, Bruce D. Lee, Nikolai Matni, George J. Pappas

Domain randomization (DR) enables sim-to-real transfer by training controllers on a distribution of simulated environments, with the goal of achieving robust performance in the real world. Although DR is widely used in practice and is often solved using simple policy gradient (PG) methods, understanding of its theoretical guarantees remains limited. Toward addressing this gap, we provide the first convergence analysis of PG methods for domain-randomized linear quadratic regulation (LQR). We show that PG converges globally to the minimizer of a finite-sample approximation of the DR objective under suitable bounds on the heterogeneity of the sampled systems. We also quantify the sample-complexity associated with achieving a small performance gap between the sample-average and population-level objectives. Additionally, we propose and analyze a discount-factor annealing algorithm that obviates the need for an initial jointly stabilizing controller, which may be challenging to find. Empirical results support our theoretical findings and highlight promising directions for future work, including risk-sensitive DR formulations and stochastic PG algorithms.

☆ Effectively Controlling Reasoning Models through Thinking Intervention

Tong Wu, Chong Xiang, Jiachen T. Wang, Prateek Mittal

Reasoning-enhanced large language models (LLMs) explicitly generate intermediate reasoning steps prior to generating final answers, helping the model excel in complex problem-solving. In this paper, we demonstrate that this emerging generation framework offers a unique opportunity for more fine-grained control over model behavior. We propose Thinking Intervention, a novel paradigm designed to explicitly guide the internal reasoning processes of LLMs by strategically inserting or revising specific thinking tokens. We conduct comprehensive evaluations across multiple tasks, including instruction following on IFEval, instruction hierarchy on SEP, and safety alignment on XSTest and SORRY-Bench. Our results demonstrate that Thinking Intervention significantly outperforms baseline prompting approaches, achieving up to 6.7% accuracy gains in instruction-following scenarios, 15.4% improvements in reasoning about instruction hierarchies, and a 40.0% increase in refusal rates for unsafe prompts using open-source DeepSeek R1 models. Overall, our work opens a promising new research avenue for controlling reasoning LLMs.

☆ Which LIME should I trust? Concepts, Challenges, and Solutions

Patrick Knab, Sascha Marton, Udo Schlegel, Christian Bartelt

As neural networks become dominant in essential systems, Explainable Artificial Intelligence (XAI) plays a crucial role in fostering trust and detecting potential misbehavior of opaque models. LIME (Local Interpretable Model-agnostic Explanations) is among the most prominent model-agnostic approaches, generating explanations by approximating the behavior of black-box models around specific instances. Despite its popularity, LIME faces challenges related to fidelity, stability, and applicability to domain-specific problems. Numerous adaptations and enhancements have been proposed to address these issues, but the growing number of developments can be overwhelming, complicating efforts to navigate LIME-related research. To the best of our knowledge, this is the first survey to comprehensively explore and collect LIME's foundational concepts and known limitations. We categorize and compare its various enhancements, offering a structured taxonomy based on intermediate steps and key issues. Our analysis provides a holistic overview of advancements in LIME, guiding future research and helping practitioners identify suitable approaches. Additionally, we provide a continuously updated interactive website (https://patrick-knab.github.io/which-lime-to-trust/), offering a concise and accessible overview of the survey.

comment: Accepted at the 3rd World Conference on eXplainable Artificial Intelligence (XAI 2025)

☆ Sim-and-Real Co-Training: A Simple Recipe for Vision-Based Robotic Manipulation

Abhiram Maddukuri, Zhenyu Jiang, Lawrence Yunliang Chen, Soroush Nasiriany, Yuqi Xie, Yu Fang, Wenqi Huang, Zu Wang, Zhenjia Xu, Nikita Chernyadev, Scott Reed, Ken Goldberg, Ajay Mandlekar, Linxi Fan, Yuke Zhu

Large real-world robot datasets hold great potential to train generalist robot models, but scaling real-world human data collection is time-consuming and resource-intensive. Simulation has great potential in supplementing large-scale data, especially with recent advances in generative AI and automated data generation tools that enable scalable creation of robot behavior datasets. However, training a policy solely in simulation and transferring it to the real world often demands substantial human effort to bridge the reality gap. A compelling alternative is to co-train the policy on a mixture of simulation and real-world datasets. Preliminary studies have recently shown this strategy to substantially improve the performance of a policy over one trained on a limited amount of real-world data. Nonetheless, the community lacks a systematic understanding of sim-and-real co-training and what it takes to reap the benefits of simulation data for real-robot learning. This work presents a simple yet effective recipe for utilizing simulation data to solve vision-based robotic manipulation tasks. We derive this recipe from comprehensive experiments that validate the co-training strategy on various simulation and real-world datasets. Using two domains--a robot arm and a humanoid--across diverse tasks, we demonstrate that simulation data can enhance real-world task performance by an average of 38%, even with notable differences between the simulation and real-world data. Videos and additional results can be found at https://co-training.github.io/

comment: Project website: https://co-training.github.io/

☆ SQuat: Subspace-orthogonal KV Cache Quantization

Hao Wang, Ligong Han, Kai Xu, Akash Srivastava

The key-value (KV) cache accelerates LLMs decoding by storing KV tensors from previously generated tokens. It reduces redundant computation at the cost of increased memory usage. To mitigate this overhead, existing approaches compress KV tensors into lower-bit representations; however, quantization errors can accumulate as more tokens are generated, potentially resulting in undesired outputs. In this paper, we introduce SQuat (Subspace-orthogonal KV cache quantization). It first constructs a subspace spanned by query tensors to capture the most critical task-related information. During key tensor quantization, it enforces that the difference between the (de)quantized and original keys remains orthogonal to this subspace, minimizing the impact of quantization errors on the attention mechanism's outputs. SQuat requires no model fine-tuning, no additional calibration dataset for offline learning, and is grounded in a theoretical framework we develop. Through numerical experiments, we show that our method reduces peak memory by 2.17 to 2.82, improves throughput by 2.45 to 3.60, and achieves more favorable benchmark scores than existing KV cache quantization algorithms.

☆ ORAL: Prompting Your Large-Scale LoRAs via Conditional Recurrent Diffusion

Rana Muhammad Shahroz Khan, Dongwen Tang, Pingzhi Li, Kai Wang, Tianlong Chen

Parameter generation has emerged as a novel paradigm for neural network development, offering an alternative to traditional neural network training by synthesizing high-quality model weights directly. In the context of Low-Rank Adaptation (LoRA) for evolving ($\textit{i.e.}$, constantly updated) large language models (LLMs), this approach promises efficient adaptation without costly retraining. However, existing methods face critical limitations in simultaneously achieving scalability and controllability. In this paper, we introduce $\texttt{ORAL}$, a novel $\textbf{conditional recurrent diffusion}$ framework that addresses these challenges. $\texttt{ORAL}$ incorporates a novel conditioning mechanism that integrates model architecture and textual task specifications, enabling the generation of task-specific LoRA parameters that can seamlessly transfer across evolving foundation models. Our approach successfully scales to billions-of-parameter LLMs and maintains controllability. Through extensive experiments across seven language tasks, four vision tasks, and three multimodal tasks using five pre-trained LLMs, we demonstrate that $\texttt{ORAL}$ generates high-quality LoRA parameters that achieve comparable or superior performance to vanilla trained counterparts.

☆ Faster Rates for No-Regret Learning in General Games via Cautious Optimism STOC 2025

Ashkan Soleymani, Georgios Piliouras, Gabriele Farina

We establish the first uncoupled learning algorithm that attains $O(n \log^2 d \log T)$ per-player regret in multi-player general-sum games, where $n$ is the number of players, $d$ is the number of actions available to each player, and $T$ is the number of repetitions of the game. Our results exponentially improve the dependence on $d$ compared to the $O(n\, d \log T)$ regret attainable by Log-Regularized Lifted Optimistic FTRL [Far+22c], and also reduce the dependence on the number of iterations $T$ from $\log^4 T$ to $\log T$ compared to Optimistic Hedge, the previously well-studied algorithm with $O(n \log d \log^4 T)$ regret [DFG21]. Our algorithm is obtained by combining the classic Optimistic Multiplicative Weights Update (OMWU) with an adaptive, non-monotonic learning rate that paces the learning process of the players, making them more cautious when their regret becomes too negative.

comment: Appeared at STOC 2025

☆ Contextual Preference Collaborative Measure Framework Based on Belief System

Hang Yu, Wei Wei, Zheng Tan, Jing-lei Liu

To reduce the human intervention in the preference measure process,this article proposes a preference collaborative measure framework based on an updated belief system,which is also capable of improving the accuracy and efficiency of preferen-ce measure algorithms.Firstly,the distance of rules and the average internal distance of rulesets are proposed for specifying the relationship between the rules.For discovering the most representative preferences that are common in all users,namely common preference,a algorithm based on average internal distance of ruleset,PRA algorithm,is proposed,which aims to finish the discoveryprocess with minimum information loss rate.Furthermore,the concept of Common belief is proposed to update the belief system,and the common preferences are the evidences of updated belief system.Then,under the belief system,the proposed belief degree and deviation degree are used to determine whether a rule confirms the belief system or not and classify the preference rules into two kinds(generalized or personalized),and eventually filters out Top-K interesting rules relying on belief degree and deviation degree.Based on above,a scalable interestingness calculation framework that can apply various formulas is proposed for accurately calculating interestingness in different conditions.At last,IMCos algorithm and IMCov algorithm are proposed as exemplars to verify the accuracy and efficiency of the framework by using weighted cosine similarity and correlation coefficients as belief degree.In experiments,the proposed algorithms are compared to two state-of-the-art algorithms and the results show that IMCos and IMCov outperform than the other two in most aspects.

comment: in Chinese language

☆ Self-Supervised Pretraining for Aerial Road Extraction

Rupert Polley, Sai Vignesh Abishek Deenadayalan, J. Marius Zöllner

Deep neural networks for aerial image segmentation require large amounts of labeled data, but high-quality aerial datasets with precise annotations are scarce and costly to produce. To address this limitation, we propose a self-supervised pretraining method that improves segmentation performance while reducing reliance on labeled data. Our approach uses inpainting-based pretraining, where the model learns to reconstruct missing regions in aerial images, capturing their inherent structure before being fine-tuned for road extraction. This method improves generalization, enhances robustness to domain shifts, and is invariant to model architecture and dataset choice. Experiments show that our pretraining significantly boosts segmentation accuracy, especially in low-data regimes, making it a scalable solution for aerial image analysis.

☆ NoProp: Training Neural Networks without Back-propagation or Forward-propagation

Qinyu Li, Yee Whye Teh, Razvan Pascanu

The canonical deep learning approach for learning requires computing a gradient term at each layer by back-propagating the error signal from the output towards each learnable parameter. Given the stacked structure of neural networks, where each layer builds on the representation of the layer below, this approach leads to hierarchical representations. More abstract features live on the top layers of the model, while features on lower layers are expected to be less abstract. In contrast to this, we introduce a new learning method named NoProp, which does not rely on either forward or backwards propagation. Instead, NoProp takes inspiration from diffusion and flow matching methods, where each layer independently learns to denoise a noisy target. We believe this work takes a first step towards introducing a new family of gradient-free learning methods, that does not learn hierarchical representations -- at least not in the usual sense. NoProp needs to fix the representation at each layer beforehand to a noised version of the target, learning a local denoising process that can then be exploited at inference. We demonstrate the effectiveness of our method on MNIST, CIFAR-10, and CIFAR-100 image classification benchmarks. Our results show that NoProp is a viable learning algorithm which achieves superior accuracy, is easier to use and computationally more efficient compared to other existing back-propagation-free methods. By departing from the traditional gradient based learning paradigm, NoProp alters how credit assignment is done within the network, enabling more efficient distributed learning as well as potentially impacting other characteristics of the learning process.

☆ Sample-Optimal Private Regression in Polynomial Time

Prashanti Anderson, Ainesh Bakshi, Mahbod Majid, Stefan Tiegel

We consider the task of privately obtaining prediction error guarantees in ordinary least-squares regression problems with Gaussian covariates (with unknown covariance structure). We provide the first sample-optimal polynomial time algorithm for this task under both pure and approximate differential privacy. We show that any improvement to the sample complexity of our algorithm would violate either statistical-query or information-theoretic lower bounds. Additionally, our algorithm is robust to a small fraction of arbitrary outliers and achieves optimal error rates as a function of the fraction of outliers. In contrast, all prior efficient algorithms either incurred sample complexities with sub-optimal dimension dependence, scaling with the condition number of the covariates, or obtained a polynomially worse dependence on the privacy parameters. Our technical contributions are two-fold: first, we leverage resilience guarantees of Gaussians within the sum-of-squares framework. As a consequence, we obtain efficient sum-of-squares algorithms for regression with optimal robustness rates and sample complexity. Second, we generalize the recent robustness-to-privacy framework [HKMN23, (arXiv:2212.05015)] to account for the geometry induced by the covariance of the input samples. This framework crucially relies on the robust estimators to be sum-of-squares algorithms, and combining the two steps yields a sample-optimal private regression algorithm. We believe our techniques are of independent interest, and we demonstrate this by obtaining an efficient algorithm for covariance-aware mean estimation, with an optimal dependence on the privacy parameters.

☆ A Systematic Evaluation of LLM Strategies for Mental Health Text Analysis: Fine-tuning vs. Prompt Engineering vs. RAG

Arshia Kermani, Veronica Perez-Rosas, Vangelis Metsis

This study presents a systematic comparison of three approaches for the analysis of mental health text using large language models (LLMs): prompt engineering, retrieval augmented generation (RAG), and fine-tuning. Using LLaMA 3, we evaluate these approaches on emotion classification and mental health condition detection tasks across two datasets. Fine-tuning achieves the highest accuracy (91% for emotion classification, 80% for mental health conditions) but requires substantial computational resources and large training sets, while prompt engineering and RAG offer more flexible deployment with moderate performance (40-68% accuracy). Our findings provide practical insights for implementing LLM-based solutions in mental health applications, highlighting the trade-offs between accuracy, computational requirements, and deployment flexibility.

☆ Evaluating machine learning models for predicting pesticides toxicity to honey bees

Jakub Adamczyk, Jakub Poziemski, Pawel Siedlecki

Small molecules play a critical role in the biomedical, environmental, and agrochemical domains, each with distinct physicochemical requirements and success criteria. Although biomedical research benefits from extensive datasets and established benchmarks, agrochemical data remain scarce, particularly with respect to species-specific toxicity. This work focuses on ApisTox, the most comprehensive dataset of experimentally validated chemical toxicity to the honey bee (\textit{Apis mellifera}), an ecologically vital pollinator. We evaluate ApisTox using a diverse suite of machine learning approaches, including molecular fingerprints, graph kernels, and graph neural networks, as well as pretrained models. Comparative analysis with medicinal datasets from the MoleculeNet benchmark reveals that ApisTox represents a distinct chemical space. Performance degradation on non-medicinal datasets, such as ApisTox, demonstrates their limited generalizability of current state-of-the-art algorithms trained solely on biomedical data. Our study highlights the need for more diverse datasets and for targeted model development geared toward the agrochemical domain.

☆ Solving the Best Subset Selection Problem via Suboptimal Algorithms

Vikram Singh, Min Sun

Best subset selection in linear regression is well known to be nonconvex and computationally challenging to solve, as the number of possible subsets grows rapidly with increasing dimensionality of the problem. As a result, finding the global optimal solution via an exact optimization method for a problem with dimensions of 1000s may take an impractical amount of CPU time. This suggests the importance of finding suboptimal procedures that can provide good approximate solutions using much less computational effort than exact methods. In this work, we introduce a new procedure and compare it with other popular suboptimal algorithms to solve the best subset selection problem. Extensive computational experiments using synthetic and real data have been performed. The results provide insights into the performance of these methods in different data settings. The new procedure is observed to be a competitive suboptimal algorithm for solving the best subset selection problem for high-dimensional data.

☆ Fair Dynamic Spectrum Access via Fully Decentralized Multi-Agent Reinforcement Learning

Yubo Zhang, Pedro Botelho, Trevor Gordon, Gil Zussman, Igor Kadota

We consider a decentralized wireless network with several source-destination pairs sharing a limited number of orthogonal frequency bands. Sources learn to adapt their transmissions (specifically, their band selection strategy) over time, in a decentralized manner, without sharing information with each other. Sources can only observe the outcome of their own transmissions (i.e., success or collision), having no prior knowledge of the network size or of the transmission strategy of other sources. The goal of each source is to maximize their own throughput while striving for network-wide fairness. We propose a novel fully decentralized Reinforcement Learning (RL)-based solution that achieves fairness without coordination. The proposed Fair Share RL (FSRL) solution combines: (i) state augmentation with a semi-adaptive time reference; (ii) an architecture that leverages risk control and time difference likelihood; and (iii) a fairness-driven reward structure. We evaluate FSRL in more than 50 network settings with different number of agents, different amounts of available spectrum, in the presence of jammers, and in an ad-hoc setting. Simulation results suggest that, when we compare FSRL with a common baseline RL algorithm from the literature, FSRL can be up to 89.0% fairer (as measured by Jain's fairness index) in stringent settings with several sources and a single frequency band, and 48.1% fairer on average.

comment: To appear in WiOpt 2025

☆ Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Model

Jingcheng Hu, Yinmin Zhang, Qi Han, Daxin Jiang, Xiangyu Zhang, Heung-Yeung Shum

We introduce Open-Reasoner-Zero, the first open source implementation of large-scale reasoning-oriented RL training focusing on scalability, simplicity and accessibility. Through extensive experiments, we demonstrate that a minimalist approach, vanilla PPO with GAE ($\lambda=1$, $\gamma=1$) and straightforward rule-based rewards, without any KL regularization, is sufficient to scale up both response length and benchmark performance, similar to the phenomenon observed in DeepSeek-R1-Zero. Using the same base model as DeepSeek-R1-Zero-Qwen-32B, our implementation achieves superior performance on AIME2024, MATH500, and the GPQA Diamond benchmark while demonstrating remarkable efficiency -- requiring only a tenth of the training steps, compared to DeepSeek-R1-Zero pipeline. In the spirit of open source, we release our source code, parameter settings, training data, and model weights across various sizes.

☆ Value of Information-based Deceptive Path Planning Under Adversarial Interventions

Wesley A. Suttle, Jesse Milzman, Mustafa O. Karabag, Brian M. Sadler, Ufuk Topcu

Existing methods for deceptive path planning (DPP) address the problem of designing paths that conceal their true goal from a passive, external observer. Such methods do not apply to problems where the observer has the ability to perform adversarial interventions to impede the path planning agent. In this paper, we propose a novel Markov decision process (MDP)-based model for the DPP problem under adversarial interventions and develop new value of information (VoI) objectives to guide the design of DPP policies. Using the VoI objectives we propose, path planning agents deceive the adversarial observer into choosing suboptimal interventions by selecting trajectories that are of low informational value to the observer. Leveraging connections to the linear programming theory for MDPs, we derive computationally efficient solution methods for synthesizing policies for performing DPP under adversarial interventions. In our experiments, we illustrate the effectiveness of the proposed solution method in achieving deceptiveness under adversarial interventions and demonstrate the superior performance of our approach to both existing DPP methods and conservative path planning approaches on illustrative gridworld problems.

comment: 10 pages, 4 figures

☆ Evaluating and Designing Sparse Autoencoders by Approximating Quasi-Orthogonality

Sewoong Lee, Adam Davies, Marc E. Canby, Julia Hockenmaier

Sparse autoencoders (SAEs) have emerged as a workhorse of modern mechanistic interpretability, but leading SAE approaches with top-$k$ style activation functions lack theoretical grounding for selecting the hyperparameter $k$. SAEs are based on the linear representation hypothesis (LRH), which assumes that the representations of large language models (LLMs) are linearly encoded, and the superposition hypothesis (SH), which states that there can be more features in the model than its dimensionality. We show that, based on the formal definitions of the LRH and SH, the magnitude of sparse feature vectors (the latent representations learned by SAEs of the dense embeddings of LLMs) can be approximated using their corresponding dense vector with a closed-form error bound. To visualize this, we propose the ZF plot, which reveals a previously unknown relationship between LLM hidden embeddings and SAE feature vectors, allowing us to make the first empirical measurement of the extent to which feature vectors of pre-trained SAEs are over- or under-activated for a given input. Correspondingly, we introduce Approximate Feature Activation (AFA), which approximates the magnitude of the ground-truth sparse feature vector, and propose a new evaluation metric derived from AFA to assess the alignment between inputs and activations. We also leverage AFA to introduce a novel SAE architecture, the top-AFA SAE, leading to SAEs that: (a) are more in line with theoretical justifications; and (b) obviate the need to tune SAE sparsity hyperparameters. Finally, we empirically demonstrate that top-AFA SAEs achieve reconstruction loss comparable to that of state-of-the-art top-k SAEs, without requiring the hyperparameter $k$ to be tuned. Our code is available at: https://github.com/SewoongLee/top-afa-sae.

☆ Learning Velocity and Acceleration: Self-Supervised Motion Consistency for Pedestrian Trajectory Prediction

Yizhou Huang, Yihua Cheng, Kezhi Wang

Understanding human motion is crucial for accurate pedestrian trajectory prediction. Conventional methods typically rely on supervised learning, where ground-truth labels are directly optimized against predicted trajectories. This amplifies the limitations caused by long-tailed data distributions, making it difficult for the model to capture abnormal behaviors. In this work, we propose a self-supervised pedestrian trajectory prediction framework that explicitly models position, velocity, and acceleration. We leverage velocity and acceleration information to enhance position prediction through feature injection and a self-supervised motion consistency mechanism. Our model hierarchically injects velocity features into the position stream. Acceleration features are injected into the velocity stream. This enables the model to predict position, velocity, and acceleration jointly. From the predicted position, we compute corresponding pseudo velocity and acceleration, allowing the model to learn from data-generated pseudo labels and thus achieve self-supervised learning. We further design a motion consistency evaluation strategy grounded in physical principles; it selects the most reasonable predicted motion trend by comparing it with historical dynamics and uses this trend to guide and constrain trajectory generation. We conduct experiments on the ETH-UCY and Stanford Drone datasets, demonstrating that our method achieves state-of-the-art performance on both datasets.

☆ Enhancing Image Resolution of Solar Magnetograms: A Latent Diffusion Model Approach

Francesco Pio Ramunno, Paolo Massa, Vitaliy Kinakh, Brandon Panos, André Csillaghy, Slava Voloshynovskiy

The spatial properties of the solar magnetic field are crucial to decoding the physical processes in the solar interior and their interplanetary effects. However, observations from older instruments, such as the Michelson Doppler Imager (MDI), have limited spatial or temporal resolution, which hinders the ability to study small-scale solar features in detail. Super resolving these older datasets is essential for uniform analysis across different solar cycles, enabling better characterization of solar flares, active regions, and magnetic network dynamics. In this work, we introduce a novel diffusion model approach for Super-Resolution and we apply it to MDI magnetograms to match the higher-resolution capabilities of the Helioseismic and Magnetic Imager (HMI). By training a Latent Diffusion Model (LDM) with residuals on downscaled HMI data and fine-tuning it with paired MDI/HMI data, we can enhance the resolution of MDI observations from 2"/pixel to 0.5"/pixel. We evaluate the quality of the reconstructed images by means of classical metrics (e.g., PSNR, SSIM, FID and LPIPS) and we check if physical properties, such as the unsigned magnetic flux or the size of an active region, are preserved. We compare our model with different variations of LDM and Denoising Diffusion Probabilistic models (DDPMs), but also with two deterministic architectures already used in the past for performing the Super-Resolution task. Furthermore, we show with an analysis in the Fourier domain that the LDM with residuals can resolve features smaller than 2", and due to the probabilistic nature of the LDM, we can asses their reliability, in contrast with the deterministic models. Future studies aim to super-resolve the temporal scale of the solar MDI instrument so that we can also have a better overview of the dynamics of the old events.

comment: Accepted for publication on A&A

☆ New Statistical Framework for Extreme Error Probability in High-Stakes Domains for Reliable Machine Learning

Umberto Michelucci, Francesca Venturini

Machine learning is vital in high-stakes domains, yet conventional validation methods rely on averaging metrics like mean squared error (MSE) or mean absolute error (MAE), which fail to quantify extreme errors. Worst-case prediction failures can have substantial consequences, but current frameworks lack statistical foundations for assessing their probability. In this work a new statistical framework, based on Extreme Value Theory (EVT), is presented that provides a rigorous approach to estimating worst-case failures. Applying EVT to synthetic and real-world datasets, this method is shown to enable robust estimation of catastrophic failure probabilities, overcoming the fundamental limitations of standard cross-validation. This work establishes EVT as a fundamental tool for assessing model reliability, ensuring safer AI deployment in new technologies where uncertainty quantification is central to decision-making or scientific analysis.

☆ Advances in Continual Graph Learning for Anti-Money Laundering Systems: A Comprehensive Review

Bruno Deprez, Wei Wei, Wouter Verbeke, Bart Baesens, Kevin Mets, Tim Verdonck

Financial institutions are required by regulation to report suspicious financial transactions related to money laundering. Therefore, they need to constantly monitor vast amounts of incoming and outgoing transactions. A particular challenge in detecting money laundering is that money launderers continuously adapt their tactics to evade detection. Hence, detection methods need constant fine-tuning. Traditional machine learning models suffer from catastrophic forgetting when fine-tuning the model on new data, thereby limiting their effectiveness in dynamic environments. Continual learning methods may address this issue and enhance current anti-money laundering (AML) practices, by allowing models to incorporate new information while retaining prior knowledge. Research on continual graph learning for AML, however, is still scarce. In this review, we critically evaluate state-of-the-art continual graph learning approaches for AML applications. We categorise methods into replay-based, regularization-based, and architecture-based strategies within the graph neural network (GNN) framework, and we provide in-depth experimental evaluations on both synthetic and real-world AML data sets that showcase the effect of the different hyperparameters. Our analysis demonstrates that continual learning improves model adaptability and robustness in the face of extreme class imbalances and evolving fraud patterns. Finally, we outline key challenges and propose directions for future research.

☆ Spatio-temporal Prediction of Fine-Grained Origin-Destination Matrices with Applications in Ridesharing

Run Yang, Runpeng Dai, Siran Gao, Xiaocheng Tang, Fan Zhou, Hongtu Zhu

Accurate spatial-temporal prediction of network-based travelers' requests is crucial for the effective policy design of ridesharing platforms. Having knowledge of the total demand between various locations in the upcoming time slots enables platforms to proactively prepare adequate supplies, thereby increasing the likelihood of fulfilling travelers' requests and redistributing idle drivers to areas with high potential demand to optimize the global supply-demand equilibrium. This paper delves into the prediction of Origin-Destination (OD) demands at a fine-grained spatial level, especially when confronted with an expansive set of local regions. While this task holds immense practical value, it remains relatively unexplored within the research community. To fill this gap, we introduce a novel prediction model called OD-CED, which comprises an unsupervised space coarsening technique to alleviate data sparsity and an encoder-decoder architecture to capture both semantic and geographic dependencies. Through practical experimentation, OD-CED has demonstrated remarkable results. It achieved an impressive reduction of up to 45% reduction in root-mean-square error and 60% in weighted mean absolute percentage error over traditional statistical methods when dealing with OD matrices exhibiting a sparsity exceeding 90%.

☆ GPU-centric Communication Schemes for HPC and ML Applications

Naveen Namashivayam

Compute nodes on modern heterogeneous supercomputing systems comprise CPUs, GPUs, and high-speed network interconnects (NICs). Parallelization is identified as a technique for effectively utilizing these systems to execute scalable simulation and deep learning workloads. The resulting inter-process communication from the distributed execution of these parallel workloads is one of the key factors contributing to its performance bottleneck. Most programming models and runtime systems enabling the communication requirements on these systems support GPU-aware communication schemes that move the GPU-attached communication buffers in the application directly from the GPU to the NIC without staging through the host memory. A CPU thread is required to orchestrate the communication operations even with support for such GPU-awareness. This survey discusses various available GPU-centric communication schemes that move the control path of the communication operations from the CPU to the GPU. This work presents the need for the new communication schemes, various GPU and NIC capabilities required to implement the schemes, and the potential use-cases addressed. Based on these discussions, challenges involved in supporting the exhibited GPU-centric communication schemes are discussed.

comment: A surveyor on Communication Schemes for Distributed HPC and ML Applications. Article in consideration for journal publication

☆ MB-ORES: A Multi-Branch Object Reasoner for Visual Grounding in Remote Sensing

Karim Radouane, Hanane Azzag, Mustapha lebbah

We propose a unified framework that integrates object detection (OD) and visual grounding (VG) for remote sensing (RS) imagery. To support conventional OD and establish an intuitive prior for VG task, we fine-tune an open-set object detector using referring expression data, framing it as a partially supervised OD task. In the first stage, we construct a graph representation of each image, comprising object queries, class embeddings, and proposal locations. Then, our task-aware architecture processes this graph to perform the VG task. The model consists of: (i) a multi-branch network that integrates spatial, visual, and categorical features to generate task-aware proposals, and (ii) an object reasoning network that assigns probabilities across proposals, followed by a soft selection mechanism for final referring object localization. Our model demonstrates superior performance on the OPT-RSVG and DIOR-RSVG datasets, achieving significant improvements over state-of-the-art methods while retaining classical OD capabilities. The code will be available in our repository: \url{https://github.com/rd20karim/MB-ORES}.

☆ Data-driven construction of a generalized kinetic collision operator from molecular dynamics

Yue Zhao, William Burby, Andrew Christlieb, Huan Lei

We introduce a data-driven approach to learn a generalized kinetic collision operator directly from molecular dynamics. Unlike the conventional (e.g., Landau) models, the present operator takes an anisotropic form that accounts for a second energy transfer arising from the collective interactions between the pair of collision particles and the environment. Numerical results show that preserving the broadly overlooked anisotropic nature of the collision energy transfer is crucial for predicting the plasma kinetics with non-negligible correlations, where the Landau model shows limitations.

☆ A Comparison of Parametric Dynamic Mode Decomposition Algorithms for Thermal-Hydraulics Applications

Stefano Riva, Andrea Missaglia, Carolina Introini, In Cheol Bang, Antonio Cammi

In recent years, algorithms aiming at learning models from available data have become quite popular due to two factors: 1) the significant developments in Artificial Intelligence techniques and 2) the availability of large amounts of data. Nevertheless, this topic has already been addressed by methodologies belonging to the Reduced Order Modelling framework, of which perhaps the most famous equation-free technique is Dynamic Mode Decomposition. This algorithm aims to learn the best linear model that represents the physical phenomena described by a time series dataset: its output is a best state operator of the underlying dynamical system that can be used, in principle, to advance the original dataset in time even beyond its span. However, in its standard formulation, this technique cannot deal with parametric time series, meaning that a different linear model has to be derived for each parameter realization. Research on this is ongoing, and some versions of a parametric Dynamic Mode Decomposition already exist. This work contributes to this research field by comparing the different algorithms presently deployed and assessing their advantages and shortcomings compared to each other. To this aim, three different thermal-hydraulics problems are considered: two benchmark 'flow over cylinder' test cases at diverse Reynolds numbers, whose datasets are, respectively, obtained with the FEniCS finite element solver and retrieved from the CFDbench dataset, and the DYNASTY experimental facility operating at Politecnico di Milano, which studies the natural circulation established by internally heated fluids for Generation IV nuclear applications, whose dataset was generated using the RELAP5 nodal solver.

☆ Many-to-Many Matching via Sparsity Controlled Optimal Transport

Weijie Liu, Han Bao, Makoto Yamada, Zenan Huang, Nenggan Zheng, Hui Qian

Many-to-many matching seeks to match multiple points in one set and multiple points in another set, which is a basis for a wide range of data mining problems. It can be naturally recast in the framework of Optimal Transport (OT). However, existing OT methods either lack the ability to accomplish many-to-many matching or necessitate careful tuning of a regularization parameter to achieve satisfactory results. This paper proposes a novel many-to-many matching method to explicitly encode many-to-many constraints while preventing the degeneration into one-to-one matching. The proposed method consists of the following two components. The first component is the matching budget constraints on each row and column of a transport plan, which specify how many points can be matched to a point at most. The second component is the deformed $q$-entropy regularization, which encourages a point to meet the matching budget maximally. While the deformed $q$-entropy was initially proposed to sparsify a transport plan, we employ it to avoid the degeneration into one-to-one matching. We optimize the objective via a penalty algorithm, which is efficient and theoretically guaranteed to converge. Experimental results on various tasks demonstrate that the proposed method achieves good performance by gleaning meaningful many-to-many matchings.

☆ Traffic Engineering in Large-scale Networks with Generalizable Graph Neural Networks

Fangtong Zhou, Xiaorui Liu, Ruozhou Yu, Guoliang Xue

Traffic engineering (TE) in large-scale computer networks has become a fundamental yet challenging problem, owing to the swift growth of global-scale cloud wide-area networks or backbone low-Earth-orbit satellite constellations. To address the scalability issue of traditional TE algorithms, learning-based approaches have been proposed, showing potential of significant efficiency improvement over state-of-the-art methods. Nevertheless, the intrinsic limitations of existing learning-based methods hinder their practical application: they are not generalizable across diverse topologies and network conditions, incur excessive training overhead, and do not respect link capacities by default. This paper proposes TELGEN, a novel TE algorithm that learns to solve TE problems efficiently in large-scale networks, while achieving superior generalizability across diverse network conditions. TELGEN is based on the novel idea of transforming the problem of "predicting the optimal TE solution" into "predicting the optimal TE algorithm", which enables TELGEN to learn and efficiently approximate the end-to-end solving process of classical optimal TE algorithms. The learned algorithm is agnostic to the exact network topology or traffic patterns, and can efficiently solve TE problems given arbitrary inputs and generalize well to unseen topologies and demands. We trained and evaluated TELGEN on random and real-world networks with up to 5000 nodes and 106 links. TELGEN achieved less than 3% optimality gap while ensuring feasibility in all cases, even when the test network had up to 20x more nodes than the largest in training. It also saved up to 84% solving time than classical optimal solver, and could reduce training time per epoch and solving time by 2-4 orders of magnitude than latest learning algorithms on the largest networks.

☆ NeuRaLaTeX: A machine learning library written in pure LaTeX

James A. D. Gardner, Will Rowan, William A. P. Smith

In this paper, we introduce NeuRaLaTeX, which we believe to be the first deep learning library written entirely in LaTeX. As part of your LaTeX document you can specify the architecture of a neural network and its loss functions, define how to generate or load training data, and specify training hyperparameters and experiments. When the document is compiled, the LaTeX compiler will generate or load training data, train the network, run experiments, and generate figures. This paper generates a random 100 point spiral dataset, trains a two layer MLP on it, evaluates on a different random spiral dataset, produces plots and tables of results. The paper took 48 hours to compile and the entire source code for NeuRaLaTeX is contained within the source code of the paper. We propose two new metrics: the Written In Latex (WIL) metric measures the proportion of a machine learning library that is written in pure LaTeX, while the Source Code Of Method in Source Code of Paper (SCOMISCOP) metric measures the proportion of a paper's implementation that is contained within the paper source. We are state-of-the-art for both metrics, outperforming the ResNet and Transformer papers, as well as the PyTorch and Tensorflow libraries. Source code, documentation, videos, crypto scams and an invitation to invest in the commercialisation of NeuRaLaTeX are available at https://www.neuralatex.com

☆ Ride-Sourcing Vehicle Rebalancing with Service Accessibility Guarantees via Constrained Mean-Field Reinforcement Learning

Matej Jusup, Kenan Zhang, Zhiyuan Hu, Barna Pásztor, Andreas Krause, Francesco Corman

The rapid expansion of ride-sourcing services such as Uber, Lyft, and Didi Chuxing has fundamentally reshaped urban transportation by offering flexible, on-demand mobility via mobile applications. Despite their convenience, these platforms confront significant operational challenges, particularly vehicle rebalancing - the strategic repositioning of thousands of vehicles to address spatiotemporal mismatches in supply and demand. Inadequate rebalancing results in prolonged rider waiting times, inefficient vehicle utilization, and inequitable distribution of services, leading to disparities in driver availability and income. To tackle these complexities, we introduce scalable continuous-state mean-field control (MFC) and reinforcement learning (MFRL) models that explicitly represent each vehicle's precise location and employ continuous repositioning actions guided by the distribution of other vehicles. To ensure equitable service distribution, an accessibility constraint is integrated within our optimal control formulation, balancing operational efficiency with equitable access to the service across geographic regions. Our approach acknowledges realistic conditions, including inherent stochasticity in transitions, the simultaneous occurrence of vehicle-rider matching, vehicles' rebalancing and cruising, and variability in rider behaviors. Crucially, we relax the traditional mean-field assumption of equal supply-demand volume, better reflecting practical scenarios. Extensive empirical evaluation using real-world data-driven simulation of Shenzhen demonstrates the real-time efficiency and robustness of our approach at the scale of tens of thousands of vehicles. The code is available at https://github.com/mjusup1501/mf-vehicle-rebalancing.

comment: 30 pages, 12 figures

☆ Predicting Targeted Therapy Resistance in Non-Small Cell Lung Cancer Using Multimodal Machine Learning

Peiying Hua, Andrea Olofson, Faraz Farhadi, Liesbeth Hondelink, Gregory Tsongalis, Konstantin Dragnev, Dagmar Hoegemann Savellano, Arief Suriawinata, Laura Tafe, Saeed Hassanpour

Lung cancer is the primary cause of cancer death globally, with non-small cell lung cancer (NSCLC) emerging as its most prevalent subtype. Among NSCLC patients, approximately 32.3% have mutations in the epidermal growth factor receptor (EGFR) gene. Osimertinib, a third-generation EGFR-tyrosine kinase inhibitor (TKI), has demonstrated remarkable efficacy in the treatment of NSCLC patients with activating and T790M resistance EGFR mutations. Despite its established efficacy, drug resistance poses a significant challenge for patients to fully benefit from osimertinib. The absence of a standard tool to accurately predict TKI resistance, including that of osimertinib, remains a critical obstacle. To bridge this gap, in this study, we developed an interpretable multimodal machine learning model designed to predict patient resistance to osimertinib among late-stage NSCLC patients with activating EGFR mutations, achieving a c-index of 0.82 on a multi-institutional dataset. This machine learning model harnesses readily available data routinely collected during patient visits and medical assessments to facilitate precision lung cancer management and informed treatment decisions. By integrating various data types such as histology images, next generation sequencing (NGS) data, demographics data, and clinical records, our multimodal model can generate well-informed recommendations. Our experiment results also demonstrated the superior performance of the multimodal model over single modality models (c-index 0.82 compared with 0.75 and 0.77), thus underscoring the benefit of combining multiple modalities in patient outcome prediction.

☆ LLM4FS: Leveraging Large Language Models for Feature Selection and How to Improve It

Jianhao Li, Xianchao Xiu

Recent advances in large language models (LLMs) have provided new opportunities for decision-making, particularly in the task of automated feature selection. In this paper, we first comprehensively evaluate LLM-based feature selection methods, covering the state-of-the-art DeepSeek-R1, GPT-o3-mini, and GPT-4.5. Then, we propose a novel hybrid strategy called LLM4FS that integrates LLMs with traditional data-driven methods. Specifically, input data samples into LLMs, and directly call traditional data-driven techniques such as random forest and forward sequential selection. Notably, our analysis reveals that the hybrid strategy leverages the contextual understanding of LLMs and the high statistical reliability of traditional data-driven methods to achieve excellent feature selection performance, even surpassing LLMs and traditional data-driven methods. Finally, we point out the limitations of its application in decision-making.

☆ Learning a Canonical Basis of Human Preferences from Binary Ratings

Kailas Vodrahalli, Wei Wei, James Zou

Recent advances in generative AI have been driven by alignment techniques such as reinforcement learning from human feedback (RLHF). RLHF and related techniques typically involve constructing a dataset of binary or ranked choice human preferences and subsequently fine-tuning models to align with these preferences. This paper shifts the focus to understanding the preferences encoded in such datasets and identifying common human preferences. We find that a small subset of 21 preference categories (selected from a set of nearly 5,000 distinct preferences) captures >89% of preference variation across individuals. This small set of preferences is analogous to a canonical basis of human preferences, similar to established findings that characterize human variation in psychology or facial recognition studies. Through both synthetic and empirical evaluations, we confirm that our low-rank, canonical set of human preferences generalizes across the entire dataset and within specific topics. We further demonstrate our preference basis' utility in model evaluation, where our preference categories offer deeper insights into model alignment, and in model training, where we show that fine-tuning on preference-defined subsets successfully aligns the model accordingly.

comment: 25 pages, 11 figures

☆ Reinforcement Learning for Safe Autonomous Two Device Navigation of Cerebral Vessels in Mechanical Thrombectomy

Harry Robertshaw, Benjamin Jackson, Jiaheng Wang, Hadi Sadati, Lennart Karstensen, Alejandro Granados, Thomas C Booth

Purpose: Autonomous systems in mechanical thrombectomy (MT) hold promise for reducing procedure times, minimizing radiation exposure, and enhancing patient safety. However, current reinforcement learning (RL) methods only reach the carotid arteries, are not generalizable to other patient vasculatures, and do not consider safety. We propose a safe dual-device RL algorithm that can navigate beyond the carotid arteries to cerebral vessels. Methods: We used the Simulation Open Framework Architecture to represent the intricacies of cerebral vessels, and a modified Soft Actor-Critic RL algorithm to learn, for the first time, the navigation of micro-catheters and micro-guidewires. We incorporate patient safety metrics into our reward function by integrating guidewire tip forces. Inverse RL is used with demonstrator data on 12 patient-specific vascular cases. Results: Our simulation demonstrates successful autonomous navigation within unseen cerebral vessels, achieving a 96% success rate, 7.0s procedure time, and 0.24 N mean forces, well below the proposed 1.5 N vessel rupture threshold. Conclusion: To the best of our knowledge, our proposed autonomous system for MT two-device navigation reaches cerebral vessels, considers safety, and is generalizable to unseen patient-specific cases for the first time. We envisage future work will extend the validation to vasculatures of different complexity and on in vitro models. While our contributions pave the way towards deploying agents in clinical settings, safety and trustworthiness will be crucial elements to consider when proposing new methodology.

☆ Graph Neural Network-Based Predictive Modeling for Robotic Plaster Printing

Diego Machain Rivera, Selen Ercan Jenny, Ping Hsun Tsai, Ena Lloret-Fritschi, Luis Salamanca, Fernando Perez-Cruz, Konstantinos E. Tatsis

This work proposes a Graph Neural Network (GNN) modeling approach to predict the resulting surface from a particle based fabrication process. The latter consists of spray-based printing of cementitious plaster on a wall and is facilitated with the use of a robotic arm. The predictions are computed using the robotic arm trajectory features, such as position, velocity and direction, as well as the printing process parameters. The proposed approach, based on a particle representation of the wall domain and the end effector, allows for the adoption of a graph-based solution. The GNN model consists of an encoder-processor-decoder architecture and is trained using data from laboratory tests, while the hyperparameters are optimized by means of a Bayesian scheme. The aim of this model is to act as a simulator of the printing process, and ultimately used for the generation of the robotic arm trajectory and the optimization of the printing parameters, towards the materialization of an autonomous plastering process. The performance of the proposed model is assessed in terms of the prediction error against unseen ground truth data, which shows its generality in varied scenarios, as well as in comparison with the performance of an existing benchmark model. The results demonstrate a significant improvement over the benchmark model, with notably better performance and enhanced error scaling across prediction steps.

☆ It's a (Blind) Match! Towards Vision-Language Correspondence without Parallel Data CVPR 2025

Dominik Schnaus, Nikita Araslanov, Daniel Cremers

The platonic representation hypothesis suggests that vision and language embeddings become more homogeneous as model and dataset sizes increase. In particular, pairwise distances within each modality become more similar. This suggests that as foundation models mature, it may become possible to match vision and language embeddings in a fully unsupervised fashion, i.e. without parallel data. We present the first feasibility study, and investigate conformity of existing vision and language foundation models in the context of unsupervised, or "blind", matching. First, we formulate unsupervised matching as a quadratic assignment problem and introduce a novel heuristic that outperforms previous solvers. We also develop a technique to find optimal matching problems, for which a non-trivial match is very likely. Second, we conduct an extensive study deploying a range of vision and language models on four datasets. Our analysis reveals that for many problem instances, vision and language representations can be indeed matched without supervision. This finding opens up the exciting possibility of embedding semantic knowledge into other modalities virtually annotation-free. As a proof of concept, we showcase an unsupervised classifier, which achieves non-trivial classification accuracy without any image-text annotation.

comment: Accepted to CVPR 2025, Project page: https://dominik-schnaus.github.io/itsamatch/

☆ CTSketch: Compositional Tensor Sketching for Scalable Neurosymbolic Learning

Seewon Choi, Alaia Solko-Breslin, Rajeev Alur, Eric Wong

Many computational tasks benefit from being formulated as the composition of neural networks followed by a discrete symbolic program. The goal of neurosymbolic learning is to train the neural networks using only end-to-end input-output labels of the composite. We introduce CTSketch, a novel, scalable neurosymbolic learning algorithm. CTSketch uses two techniques to improve the scalability of neurosymbolic inference: decompose the symbolic program into sub-programs and summarize each sub-program with a sketched tensor. This strategy allows us to approximate the output distribution of the program with simple tensor operations over the input distributions and summaries. We provide theoretical insight into the maximum error of the approximation. Furthermore, we evaluate CTSketch on many benchmarks from the neurosymbolic literature, including some designed for evaluating scalability. Our results show that CTSketch pushes neurosymbolic learning to new scales that have previously been unattainable by obtaining high accuracy on tasks involving over one thousand inputs.

comment: 15 pages, 6 figures

☆ IMPACT: A Generic Semantic Loss for Multimodal Medical Image Registration

Valentin Boussot, Cédric Hémon, Jean-Claude Nunes, Jason Downling, Simon Rouzé, Caroline Lafond, Anaïs Barateau, Jean-Louis Dillenseger

Image registration is fundamental in medical imaging, enabling precise alignment of anatomical structures for diagnosis, treatment planning, image-guided treatment or longitudinal monitoring. This work introduces IMPACT (Image Metric with Pretrained model-Agnostic Comparison for Transmodality registration), a generic semantic similarity metric designed for seamless integration into diverse image registration frameworks (such as Elastix and Voxelmorph). It compares deep learning-based features extracted from medical images without requiring task-specific training, ensuring broad applicability across various modalities. By leveraging the features of the large-scale pretrained TotalSegmentator models and the ability to integrate Segment Anything Model (SAM) and other large-scale segmentation networks, this approach offers significant advantages. It provides robust, scalable, and efficient solutions for multimodal image registration. The IMPACT loss was evaluated on five challenging registration tasks involving thoracic CT/CBCT, and pelvic MR/CT datasets. Quantitative metrics, such as Target Registration Error and Dice Similarity Coefficient, demonstrated significant improvements in anatomical alignment compared to baseline methods. Qualitative analyses further confirmed the increased robustness of the proposed metric in the face of noise, artifacts, and modality variations. IMPACT's versatility and efficiency make it a valuable tool for advancing registration performance in clinical and research applications, addressing critical challenges in multimodal medical imaging.

comment: Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). This is a preprint version and has not been peer-reviewed

☆ Inductive Graph Representation Learning with Quantum Graph Neural Networks

Arthur M. Faria, Ignacio F. Graña, Savvas Varsamopoulos

Quantum Graph Neural Networks (QGNNs) present a promising approach for combining quantum computing with graph-structured data processing. While classical Graph Neural Networks (GNNs) are renowned for their scalability and robustness, existing QGNNs often lack flexibility due to graph-specific quantum circuit designs, limiting their applicability to a narrower range of graph-structured problems, falling short of real-world scenarios. To address these limitations, we propose a versatile QGNN framework inspired by the classical GraphSAGE approach, utilizing quantum models as aggregators. In this work, we integrate established techniques for inductive representation learning on graphs with parametrized quantum convolutional and pooling layers, effectively bridging classical and quantum paradigms. The convolutional layer is flexible, enabling tailored designs for specific problems. Benchmarked on a node regression task with the QM9 dataset, we demonstrate that our framework successfully models a non-trivial molecular dataset, achieving performance comparable to classical GNNs. In particular, we show that our quantum approach exhibits robust generalization across molecules with varying numbers of atoms without requiring circuit modifications, slightly outperforming classical GNNs. Furthermore, we numerically investigate the scalability of the QGNN framework. Specifically, we demonstrate the absence of barren plateaus in our architecture as the number of qubits increases, suggesting that the proposed quantum model can be extended to handle larger and more complex graph-based problems effectively.

comment: 18 pages, 6 figures

☆ Level the Level: Balancing Game Levels for Asymmetric Player Archetypes With Reinforcement Learning

Florian Rupp, Kai Eckert

Balancing games, especially those with asymmetric multiplayer content, requires significant manual effort and extensive human playtesting during development. For this reason, this work focuses on generating balanced levels tailored to asymmetric player archetypes, where the disparity in abilities is balanced entirely through the level design. For instance, while one archetype may have an advantage over another, both should have an equal chance of winning. We therefore conceptualize game balancing as a procedural content generation problem and build on and extend a recently introduced method that uses reinforcement learning to balance tile-based game levels. We evaluate the method on four different player archetypes and demonstrate its ability to balance a larger proportion of levels compared to two baseline approaches. Furthermore, our results indicate that as the disparity between player archetypes increases, the required number of training steps grows, while the model's accuracy in achieving balance decreases.

comment: Accepted at the ACM International Conference on the Foundations of Digital Games (FDG) 2025

☆ New universal operator approximation theorem for encoder-decoder architectures (Preprint)

Janek Gödeke, Pascal Fernsel

Motivated by the rapidly growing field of mathematics for operator approximation with neural networks, we present a novel universal operator approximation theorem for a broad class of encoder-decoder architectures. In this study, we focus on approximating continuous operators in $\mathcal{C}(\mathcal{X}, \mathcal{Y})$, where $\mathcal{X}$ and $\mathcal{Y}$ are infinite-dimensional normed or metric spaces, and we consider uniform convergence on compact subsets of $\mathcal{X}$. Unlike standard results in the operator learning literature, we investigate the case where the approximating operator sequence can be chosen independently of the compact sets. Taking a topological perspective, we analyze different types of operator approximation and show that compact-set-independent approximation is a strictly stronger property in most relevant operator learning frameworks. To establish our results, we introduce a new approximation property tailored to encoder-decoder architectures, which enables us to prove a universal operator approximation theorem ensuring uniform convergence on every compact subset. This result unifies and extends existing universal operator approximation theorems for various encoder-decoder architectures, including classical DeepONets, BasisONets, special cases of MIONets, architectures based on frames and other related approaches.

comment: 34 pages

☆ Controlled Latent Diffusion Models for 3D Porous Media Reconstruction

Danilo Naiff, Bernardo P. Schaeffer, Gustavo Pires, Dragan Stojkovic, Thomas Rapstine, Fabio Ramos

Three-dimensional digital reconstruction of porous media presents a fundamental challenge in geoscience, requiring simultaneous resolution of fine-scale pore structures while capturing representative elementary volumes. We introduce a computational framework that addresses this challenge through latent diffusion models operating within the EDM framework. Our approach reduces dimensionality via a custom variational autoencoder trained in binary geological volumes, improving efficiency and also enabling the generation of larger volumes than previously possible with diffusion models. A key innovation is our controlled unconditional sampling methodology, which enhances distribution coverage by first sampling target statistics from their empirical distributions, then generating samples conditioned on these values. Extensive testing on four distinct rock types demonstrates that conditioning on porosity - a readily computable statistic - is sufficient to ensure a consistent representation of multiple complex properties, including permeability, two-point correlation functions, and pore size distributions. The framework achieves better generation quality than pixel-space diffusion while enabling significantly larger volume reconstruction (256-cube voxels) with substantially reduced computational requirements, establishing a new state-of-the-art for digital rock physics applications.

comment: 58 pages

☆ Riemannian Multiplicative Update for Sparse Simplex constraint using oblique rotation manifold

Flavia Esposito, Andersen Ang

We propose a new manifold optimization method to solve low-rank problems with sparse simplex constraints (variables are simultaneous nonnegativity, sparsity, and sum-to-1) that are beneficial in applications. The proposed approach exploits oblique rotation manifolds, rewrite the problem, and introduce a new Riemannian optimization method. Experiments on synthetic datasets compared to the standard Euclidean method show the effectiveness of the proposed method.

comment: 8 pages, 1 figure

☆ Physics-informed neural networks for hidden boundary detection and flow field reconstruction

Yongzheng Zhu, Weizheng Chen, Jian Deng, Xin Bian

Simultaneously detecting hidden solid boundaries and reconstructing flow fields from sparse observations poses a significant inverse challenge in fluid mechanics. This study presents a physics-informed neural network (PINN) framework designed to infer the presence, shape, and motion of static or moving solid boundaries within a flow field. By integrating a body fraction parameter into the governing equations, the model enforces no-slip/no-penetration boundary conditions in solid regions while preserving conservation laws of fluid dynamics. Using partial flow field data, the method simultaneously reconstructs the unknown flow field and infers the body fraction distribution, thereby revealing solid boundaries. The framework is validated across diverse scenarios, including incompressible Navier-Stokes and compressible Euler flows, such as steady flow past a fixed cylinder, an inline oscillating cylinder, and subsonic flow over an airfoil. The results demonstrate accurate detection of hidden boundaries, reconstruction of missing flow data, and estimation of trajectories and velocities of a moving body. Further analysis examines the effects of data sparsity, velocity-only measurements, and noise on inference accuracy. The proposed method exhibits robustness and versatility, highlighting its potential for applications when only limited experimental or numerical data are available.

comment: 21 pages, 17 figures

☆ From Colors to Classes: Emergence of Concepts in Vision Transformers

Teresa Dorszewski, Lenka Tětková, Robert Jenssen, Lars Kai Hansen, Kristoffer Knutsen Wickstrøm

Vision Transformers (ViTs) are increasingly utilized in various computer vision tasks due to their powerful representation capabilities. However, it remains understudied how ViTs process information layer by layer. Numerous studies have shown that convolutional neural networks (CNNs) extract features of increasing complexity throughout their layers, which is crucial for tasks like domain adaptation and transfer learning. ViTs, lacking the same inductive biases as CNNs, can potentially learn global dependencies from the first layers due to their attention mechanisms. Given the increasing importance of ViTs in computer vision, there is a need to improve the layer-wise understanding of ViTs. In this work, we present a novel, layer-wise analysis of concepts encoded in state-of-the-art ViTs using neuron labeling. Our findings reveal that ViTs encode concepts with increasing complexity throughout the network. Early layers primarily encode basic features such as colors and textures, while later layers represent more specific classes, including objects and animals. As the complexity of encoded concepts increases, the number of concepts represented in each layer also rises, reflecting a more diverse and specific set of features. Additionally, different pretraining strategies influence the quantity and category of encoded concepts, with finetuning to specific downstream tasks generally reducing the number of encoded concepts and shifting the concepts to more relevant categories.

comment: Preprint. Accepted at The 3rd World Conference on eXplainable Artificial Intelligence

☆ HACTS: a Human-As-Copilot Teleoperation System for Robot Learning

Zhiyuan Xu, Yinuo Zhao, Kun Wu, Ning Liu, Junjie Ji, Zhengping Che, Chi Harold Liu, Jian Tang

Teleoperation is essential for autonomous robot learning, especially in manipulation tasks that require human demonstrations or corrections. However, most existing systems only offer unilateral robot control and lack the ability to synchronize the robot's status with the teleoperation hardware, preventing real-time, flexible intervention. In this work, we introduce HACTS (Human-As-Copilot Teleoperation System), a novel system that establishes bilateral, real-time joint synchronization between a robot arm and teleoperation hardware. This simple yet effective feedback mechanism, akin to a steering wheel in autonomous vehicles, enables the human copilot to intervene seamlessly while collecting action-correction data for future learning. Implemented using 3D-printed components and low-cost, off-the-shelf motors, HACTS is both accessible and scalable. Our experiments show that HACTS significantly enhances performance in imitation learning (IL) and reinforcement learning (RL) tasks, boosting IL recovery capabilities and data efficiency, and facilitating human-in-the-loop RL. HACTS paves the way for more effective and interactive human-robot collaboration and data-collection, advancing the capabilities of robot manipulation.

☆ TransMamba: Flexibly Switching between Transformer and Mamba

Yixing Li, Ruobing Xie, Zhen Yang, Xingwu Sun, Shuaipeng Li, Weidong Han, Zhanhui Kang, Yu Cheng, Chengzhong Xu, Di Wang, Jie Jiang

Transformers are the cornerstone of modern large language models, but their quadratic computational complexity limits efficiency in long-sequence processing. Recent advancements in Mamba, a state space model (SSM) with linear complexity, offer promising efficiency gains but suffer from unstable contextual learning and multitask generalization. This paper proposes TransMamba, a novel framework that unifies Transformer and Mamba through shared parameter matrices (e.g., QKV and CBx), and thus could dynamically switch between attention and SSM mechanisms at different token lengths and layers. We design the Memory converter to bridge Transformer and Mamba by converting attention outputs into SSM-compatible states, ensuring seamless information flow at TransPoints where the transformation happens. The TransPoint scheduling is also thoroughly explored for further improvements. We conducted extensive experiments demonstrating that TransMamba achieves superior training efficiency and performance compared to baselines, and validated the deeper consistency between Transformer and Mamba paradigms, offering a scalable solution for next-generation sequence modeling.

comment: Preprint. Under review

☆ Artificial Conversations, Real Results: Fostering Language Detection with Synthetic Data

Fatemeh Mohammadi, Tommaso Romano, Samira Maghool, Paolo Ceravolo

Collecting high-quality training data is essential for fine-tuning Large Language Models (LLMs). However, acquiring such data is often costly and time-consuming, especially for non-English languages such as Italian. Recently, researchers have begun to explore the use of LLMs to generate synthetic datasets as a viable alternative. This study proposes a pipeline for generating synthetic data and a comprehensive approach for investigating the factors that influence the validity of synthetic data generated by LLMs by examining how model performance is affected by metrics such as prompt strategy, text length and target position in a specific task, i.e. inclusive language detection in Italian job advertisements. Our results show that, in most cases and across different metrics, the fine-tuned models trained on synthetic data consistently outperformed other models on both real and synthetic test datasets. The study discusses the practical implications and limitations of using synthetic data for language detection tasks with LLMs.

☆ Accelerated Airfoil Design Using Neural Network Approaches

Anantram Patel, Nikhil Mogre, Mandar Mane, Jayavardhan Reddy Enumula, Vijay Kumar Sutrakar

In this paper, prediction of airfoil shape from targeted pressure distribution (suction and pressure sides) and vice versa is demonstrated using both Convolutional Neural Networks (CNNs) and Deep Neural Networks (DNNs) techniques. The dataset is generated for 1600 airfoil shapes, with simulations carried out at Reynolds numbers (Re) ranging from 10,000 and 90,00,000 and angles of attack (AoA) ranging from 0 to 15 degrees, ensuring the dataset captured diverse aerodynamic conditions. Five different CNN and DNN models are developed depending on the input/output parameters. Results demonstrate that the refined models exhibit improved efficiency, with the DNN model achieving a multi-fold reduction in training time compared to the CNN model for complex datasets consisting of varying airfoil, Re, and AoA. The predicted airfoil shapes/pressure distribution closely match the targeted values, validating the effectiveness of deep learning frameworks. However, the performance of CNN models is found to be better compared to DNN models. Lastly, a flying wing aircraft model of wingspan >10 m is considered for the prediction of pressure distribution along the chordwise. The proposed CNN and DNN models show promising results. This research underscores the potential of deep learning models accelerating aerodynamic optimization and advancing the design of high-performance airfoils.

☆ Frequency-Aware Attention-LSTM for PM$_{2.5}$ Time Series Forecasting

Jiahui LU, Shuang Wu, Zhenkai Qin, Dongze Wu, Guifang Yang

To enhance the accuracy and robustness of PM$_{2.5}$ concentration forecasting, this paper introduces FALNet, a Frequency-Aware LSTM Network that integrates frequency-domain decomposition, temporal modeling, and attention-based refinement. The model first applies STL and FFT to extract trend, seasonal, and denoised residual components, effectively filtering out high-frequency noise. The filtered residuals are then fed into a stacked LSTM to capture long-term dependencies, followed by a multi-head attention mechanism that dynamically focuses on key time steps. Experiments conducted on real-world urban air quality datasets demonstrate that FALNet consistently outperforms conventional models across standard metrics such as MAE, RMSE, and $R^2$. The model shows strong adaptability in capturing sharp fluctuations during pollution peaks and non-stationary conditions. These results validate the effectiveness and generalizability of FALNet for real-time air pollution prediction, environmental risk assessment, and decision-making support.

☆ AutoML Algorithms for Online Generalized Additive Model Selection: Application to Electricity Demand Forecasting

Keshav Das, Julie Keisler, Margaux Brégère, Amaury Durand

Electricity demand forecasting is key to ensuring that supply meets demand lest the grid would blackout. Reliable short-term forecasts may be obtained by combining a Generalized Additive Models (GAM) with a State-Space model (Obst et al., 2021), leading to an adaptive (or online) model. A GAM is an over-parameterized linear model defined by a formula and a state-space model involves hyperparameters. Both the formula and adaptation parameters have to be fixed before model training and have a huge impact on the model's predictive performance. We propose optimizing them using the DRAGON package of Keisler (2025), originally designed for neural architecture search. This work generalizes it for automated online generalized additive model selection by defining an efficient modeling of the search space (namely, the space of the GAM formulae and adaptation parameters). Its application to short-term French electricity demand forecasting demonstrates the relevance of the approach

comment: 13 pages, 1 figure

☆ Crossmodal Knowledge Distillation with WordNet-Relaxed Text Embeddings for Robust Image Classification

Chenqi Guo, Mengshuo Rong, Qianli Feng, Rongfan Feng, Yinglong Ma

Crossmodal knowledge distillation (KD) aims to enhance a unimodal student using a multimodal teacher model. In particular, when the teacher's modalities include the student's, additional complementary information can be exploited to improve knowledge transfer. In supervised image classification, image datasets typically include class labels that represent high-level concepts, suggesting a natural avenue to incorporate textual cues for crossmodal KD. However, these labels rarely capture the deeper semantic structures in real-world visuals and can lead to label leakage if used directly as inputs, ultimately limiting KD performance. To address these issues, we propose a multi-teacher crossmodal KD framework that integrates CLIP image embeddings with learnable WordNet-relaxed text embeddings under a hierarchical loss. By avoiding direct use of exact class names and instead using semantically richer WordNet expansions, we mitigate label leakage and introduce more diverse textual cues. Experiments show that this strategy significantly boosts student performance, whereas noisy or overly precise text embeddings hinder distillation efficiency. Interpretability analyses confirm that WordNet-relaxed prompts encourage heavier reliance on visual features over textual shortcuts, while still effectively incorporating the newly introduced textual cues. Our method achieves state-of-the-art or second-best results on six public datasets, demonstrating its effectiveness in advancing crossmodal KD.

☆ Bayesian Predictive Coding

Alexander Tschantz, Magnus Koudahl, Hampus Linander, Lancelot Da Costa, Conor Heins, Jeff Beck, Christopher Buckley

Predictive coding (PC) is an influential theory of information processing in the brain, providing a biologically plausible alternative to backpropagation. It is motivated in terms of Bayesian inference, as hidden states and parameters are optimised via gradient descent on variational free energy. However, implementations of PC rely on maximum \textit{a posteriori} (MAP) estimates of hidden states and maximum likelihood (ML) estimates of parameters, limiting their ability to quantify epistemic uncertainty. In this work, we investigate a Bayesian extension to PC that estimates a posterior distribution over network parameters. This approach, termed Bayesian Predictive coding (BPC), preserves the locality of PC and results in closed-form Hebbian weight updates. Compared to PC, our BPC algorithm converges in fewer epochs in the full-batch setting and remains competitive in the mini-batch setting. Additionally, we demonstrate that BPC offers uncertainty quantification comparable to existing methods in Bayesian deep learning, while also improving convergence properties. Together, these results suggest that BPC provides a biologically plausible method for Bayesian learning in the brain, as well as an attractive approach to uncertainty quantification in deep learning.

☆ Tree-Guided $L_1$-Convex Clustering

Bingyuan Zhang, Yoshikazu Terada

Convex clustering is a modern clustering framework that guarantees globally optimal solutions and performs comparably to other advanced clustering methods. However, obtaining a complete dendrogram (clusterpath) for large-scale datasets remains computationally challenging due to the extensive costs associated with iterative optimization approaches. To address this limitation, we develop a novel convex clustering algorithm called Tree-Guided $L_1$-Convex Clustering (TGCC). We first focus on the fact that the loss function of $L_1$-convex clustering with tree-structured weights can be efficiently optimized using a dynamic programming approach. We then develop an efficient cluster fusion algorithm that utilizes the tree structure of the weights to accelerate the optimization process and eliminate the issue of cluster splits commonly observed in convex clustering. By combining the dynamic programming approach with the cluster fusion algorithm, the TGCC algorithm achieves superior computational efficiency without sacrificing clustering performance. Remarkably, our TGCC algorithm can construct a complete clusterpath for $10^6$ points in $\mathbb{R}^2$ within 15 seconds on a standard laptop without the need for parallel or distributed computing frameworks. Moreover, we extend the TGCC algorithm to develop biclustering and sparse convex clustering algorithms.

☆ Learning 3D-Gaussian Simulators from RGB Videos

Mikel Zhobro, Andreas René Geist, Georg Martius

Learning physics simulations from video data requires maintaining spatial and temporal consistency, a challenge often addressed with strong inductive biases or ground-truth 3D information -- limiting scalability and generalization. We introduce 3DGSim, a 3D physics simulator that learns object dynamics end-to-end from multi-view RGB videos. It encodes images into a 3D Gaussian particle representation, propagates dynamics via a transformer, and renders frames using 3D Gaussian splatting. By jointly training inverse rendering with a dynamics transformer using a temporal encoding and merging layer, 3DGSimembeds physical properties into point-wise latent vectors without enforcing explicit connectivity constraints. This enables the model to capture diverse physical behaviors, from rigid to elastic and cloth-like interactions, along with realistic lighting effects that also generalize to unseen multi-body interactions and novel scene edits.

☆ CITRAS: Covariate-Informed Transformer for Time Series Forecasting

Yosuke Yamaguchi, Issei Suemitsu, Wenpeng Wei

Covariates play an indispensable role in practical time series forecasting, offering rich context from the past and sometimes extending into the future. However, their availability varies depending on the scenario, and situations often involve multiple target variables simultaneously. Moreover, the cross-variate dependencies between them are multi-granular, with some covariates having a short-term impact on target variables and others showing long-term correlations. This heterogeneity and the intricate dependencies arising in covariate-informed forecasting present significant challenges to existing deep models. To address these issues, we propose CITRAS, a patch-based Transformer that flexibly leverages multiple targets and covariates covering both the past and the future forecasting horizon. While preserving the strong autoregressive capabilities of the canonical Transformer, CITRAS introduces two novel mechanisms in patch-wise cross-variate attention: Key-Value (KV) Shift and Attention Score Smoothing. KV Shift seamlessly incorporates future known covariates into the forecasting of target variables based on their concurrent dependencies. Additionally, Attention Score Smoothing transforms locally accurate patch-wise cross-variate dependencies into global variate-level dependencies by smoothing the past series of attention scores. Experimentally, CITRAS achieves state-of-the-art performance in both covariate-informed and multivariate forecasting, demonstrating its versatile ability to leverage cross-variate dependency for improved forecasting accuracy.

☆ Rethinking Key-Value Cache Compression Techniques for Large Language Model Serving

Wei Gao, Xinyu Zhou, Peng Sun, Tianwei Zhang, Yonggang Wen

Key-Value cache (\texttt{KV} \texttt{cache}) compression has emerged as a promising technique to optimize Large Language Model (LLM) serving. It primarily decreases the memory consumption of \texttt{KV} \texttt{cache} to reduce the computation cost. Despite the development of many compression algorithms, their applications in production environments are still not prevalent. In this paper, we revisit mainstream \texttt{KV} \texttt{cache} compression solutions from a practical perspective. Our contributions are three-fold. First, we comprehensively review existing algorithmic designs and benchmark studies for \texttt{KV} \texttt{cache} compression and identify missing pieces in their performance measurement, which could hinder their adoption in practice. Second, we empirically evaluate representative \texttt{KV} \texttt{cache} compression methods to uncover two key issues that affect the computational efficiency: (1) while compressing \texttt{KV} \texttt{cache} can reduce memory consumption, current implementations (e.g., FlashAttention, PagedAttention) do not optimize for production-level LLM serving, resulting in suboptimal throughput performance; (2) compressing \texttt{KV} \texttt{cache} may lead to longer outputs, resulting in increased end-to-end latency. We further investigate the accuracy performance of individual samples rather than the overall performance, revealing the intrinsic limitations in \texttt{KV} \texttt{cache} compression when handling specific LLM tasks. Third, we provide tools to shed light on future \texttt{KV} \texttt{cache} compression studies and facilitate their practical deployment in production. They are open-sourced in \href{https://github.com/LLMkvsys/rethink-kv-compression}{https://github.com/LLMkvsys/rethink-kv-compression}.

comment: 21 pages, 18 figures, published to MLSys2025

☆ Deep Nets as Hamiltonians

Mike Winer, Boris Hanin

Neural networks are complex functions of both their inputs and parameters. Much prior work in deep learning theory analyzes the distribution of network outputs at a fixed a set of inputs (e.g. a training dataset) over random initializations of the network parameters. The purpose of this article is to consider the opposite situation: we view a randomly initialized Multi-Layer Perceptron (MLP) as a Hamiltonian over its inputs. For typical realizations of the network parameters, we study the properties of the energy landscape induced by this Hamiltonian, focusing on the structure of near-global minimum in the limit of infinite width. Specifically, we use the replica trick to perform an exact analytic calculation giving the entropy (log volume of space) at a given energy. We further derive saddle point equations that describe the overlaps between inputs sampled iid from the Gibbs distribution induced by the random MLP. For linear activations we solve these saddle point equations exactly. But we also solve them numerically for a variety of depths and activation functions, including $\tanh, \sin, \text{ReLU}$, and shaped non-linearities. We find even at infinite width a rich range of behaviors. For some non-linearities, such as $\sin$, for instance, we find that the landscapes of random MLPs exhibit full replica symmetry breaking, while shallow $\tanh$ and ReLU networks or deep shaped MLPs are instead replica symmetric.

comment: 19+7 pages

☆ Federated Structured Sparse PCA for Anomaly Detection in IoT Networks

Chenyi Huang, Xinrong Li, Xianchao Xiu

Although federated learning has gained prominence as a privacy-preserving framework tailored for distributed Internet of Things (IoT) environments, current federated principal component analysis (PCA) methods lack integration of sparsity, a critical feature for robust anomaly detection. To address this limitation, we propose a novel federated structured sparse PCA (FedSSP) approach for anomaly detection in IoT networks. The proposed model uniquely integrates double sparsity regularization: (1) row-wise sparsity governed by $\ell_{2,p}$-norm with $p\in[0,1)$ to eliminate redundant feature dimensions, and (2) element-wise sparsity via $\ell_{q}$-norm with $q\in[0,1)$ to suppress noise-sensitive components. To efficiently solve this non-convex optimization problem in a distributed setting, we devise a proximal alternating minimization (PAM) algorithm with rigorous theoretical proofs establishing its convergence guarantees. Experiments on real datasets validate that incorporating structured sparsity enhances both model interpretability and detection accuracy.

☆ The more the merrier: logical and multistage processors in credit scoring

Arturo Pérez-Peralta, Sandra Benítez-Peña, Rosa E. Lillo

Machine Learning algorithms are ubiquitous in key decision-making contexts such as organizational justice or healthcare, which has spawned a great demand for fairness in these procedures. In this paper we focus on the application of fair ML in finance, more concretely on the use of fairness techniques on credit scoring. This paper makes two contributions. On the one hand, it addresses the existent gap concerning the application of established methods in the literature to the case of multiple sensitive variables through the use of a new technique called logical processors (LP). On the other hand, it also explores the novel method of multistage processors (MP) to investigate whether the combination of fairness methods can work synergistically to produce solutions with improved fairness or accuracy. Furthermore, we examine the intersection of these two lines of research by exploring the integration of fairness methods in the multivariate case. The results are very promising and suggest that logical processors are an appropriate way of handling multiple sensitive variables. Furthermore, multistage processors are capable of improving the performance of existing methods.

comment: 34 pages, 14 figures

☆ Noise-based reward-modulated learning

Jesús García Fernández, Nasir Ahmad, Marcel van Gerven

Recent advances in reinforcement learning (RL) have led to significant improvements in task performance. However, training neural networks in an RL regime is typically achieved in combination with backpropagation, limiting their applicability in resource-constrained environments or when using non-differentiable neural networks. While noise-based alternatives like reward-modulated Hebbian learning (RMHL) have been proposed, their performance has remained limited, especially in scenarios with delayed rewards, which require retrospective credit assignment over time. Here, we derive a novel noise-based learning rule that addresses these challenges. Our approach combines directional derivative theory with Hebbian-like updates to enable efficient, gradient-free learning in RL. It features stochastic noisy neurons which can approximate gradients, and produces local synaptic updates modulated by a global reward signal. Drawing on concepts from neuroscience, our method uses reward prediction error as its optimization target to generate increasingly advantageous behavior, and incorporates an eligibility trace to facilitate temporal credit assignment in environments with delayed rewards. Its formulation relies on local information alone, making it compatible with implementations in neuromorphic hardware. Experimental validation shows that our approach significantly outperforms RMHL and is competitive with BP-based baselines, highlighting the promise of noise-based, biologically inspired learning for low-power and real-time applications.

☆ Machine Learning-assisted High-speed Combinatorial Optimization with Ising Machines for Dynamically Changing Problems

Yohei Hamakawa, Tomoya Kashimata, Masaya Yamasaki, Kosuke Tatsumura

Quantum or quantum-inspired Ising machines have recently shown promise in solving combinatorial optimization problems in a short time. Real-world applications, such as time division multiple access (TDMA) scheduling for wireless multi-hop networks and financial trading, require solving those problems sequentially where the size and characteristics change dynamically. However, using Ising machines involves challenges to shorten system-wide latency due to the transfer of large Ising model or the cloud access and to determine the parameters for each problem. Here we show a combinatorial optimization method using embedded Ising machines, which enables solving diverse problems at high speed without runtime parameter tuning. We customize the algorithm and circuit architecture of the simulated bifurcation-based Ising machine to compress the Ising model and accelerate computation and then built a machine learning model to estimate appropriate parameters using extensive training data. In TDMA scheduling for wireless multi-hop networks, our demonstration has shown that the sophisticated system can adapt to changes in the problem and showed that it has a speed advantage over conventional methods.

☆ Green MLOps to Green GenOps: An Empirical Study of Energy Consumption in Discriminative and Generative AI Operations

Adrián Sánchez-Mompó, Ioannis Mavromatis, Peizheng Li, Konstantinos Katsaros, Aftab Khan

This study presents an empirical investigation into the energy consumption of Discriminative and Generative AI models within real-world MLOps pipelines. For Discriminative models, we examine various architectures and hyperparameters during training and inference and identify energy-efficient practices. For Generative AI, Large Language Models (LLMs) are assessed, focusing primarily on energy consumption across different model sizes and varying service requests. Our study employs software-based power measurements, ensuring ease of replication across diverse configurations, models, and datasets. We analyse multiple models and hardware setups to uncover correlations among various metrics, identifying key contributors to energy consumption. The results indicate that for Discriminative models, optimising architectures, hyperparameters, and hardware can significantly reduce energy consumption without sacrificing performance. For LLMs, energy efficiency depends on balancing model size, reasoning complexity, and request-handling capacity, as larger models do not necessarily consume more energy when utilisation remains low. This analysis provides practical guidelines for designing green and sustainable ML operations, emphasising energy consumption and carbon footprint reductions while maintaining performance. This paper can serve as a benchmark for accurately estimating total energy use across different types of AI models.

comment: Published to MDPI Information - Artificial Intelligence Section

☆ Detecting Localized Density Anomalies in Multivariate Data via Coin-Flip Statistics

Sebastian Springer, Andre Scaffidi, Maximilian Autenrieth, Gabriella Contardo, Alessandro Laio, Roberto Trotta, Heikki Haario

Detecting localized density differences in multivariate data is a crucial task in computational science. Such anomalies can indicate a critical system failure, lead to a groundbreaking scientific discovery, or reveal unexpected changes in data distribution. We introduce EagleEye, an anomaly detection method to compare two multivariate datasets with the aim of identifying local density anomalies, namely over- or under-densities affecting only localised regions of the feature space. Anomalies are detected by modelling, for each point, the ordered sequence of its neighbours' membership label as a coin-flipping process and monitoring deviations from the expected behaviour of such process. A unique advantage of our method is its ability to provide an accurate, entirely unsupervised estimate of the local signal purity. We demonstrate its effectiveness through experiments on both synthetic and real-world datasets. In synthetic data, EagleEye accurately detects anomalies in multiple dimensions even when they affect a tiny fraction of the data. When applied to a challenging resonant anomaly detection benchmark task in simulated Large Hadron Collider data, EagleEye successfully identifies particle decay events present in just 0.3% of the dataset. In global temperature data, EagleEye uncovers previously unidentified, geographically localised changes in temperature fields that occurred in the most recent years. Thanks to its key advantages of conceptual simplicity, computational efficiency, trivial parallelisation, and scalability, EagleEye is widely applicable across many fields.

☆ Model Hemorrhage and the Robustness Limits of Large Language Models

Ziyang Ma, Zuchao Li, Lefei Zhang, Gui-Song Xia, Bo Du, Liangpei Zhang, Dacheng Tao

Large language models (LLMs) demonstrate strong performance across natural language processing tasks, yet undergo significant performance degradation when modified for deployment through quantization, pruning, or decoding strategy adjustments. We define this phenomenon as model hemorrhage - performance decline caused by parameter alterations and architectural changes. Through systematic analysis of various LLM frameworks, we identify key vulnerability patterns: layer expansion frequently disrupts attention mechanisms, compression techniques induce information loss cascades, and decoding adjustments amplify prediction divergences. Our investigation reveals transformer architectures exhibit inherent robustness thresholds that determine hemorrhage severity across modification types. We propose three mitigation strategies: gradient-aware pruning preserves critical weight pathways, dynamic quantization scaling maintains activation integrity, and decoding calibration aligns generation trajectories with original model distributions. This work establishes foundational metrics for evaluating model stability during adaptation, providing practical guidelines for maintaining performance while enabling efficient LLM deployment. Our findings advance understanding of neural network resilience under architectural transformations, particularly for large-scale language models.

comment: 33 pages, 18 figures

☆ Certified Approximate Reachability (CARe): Formal Error Bounds on Deep Learning of Reachable Sets

Prashant Solanki, Nikolaus Vertovec, Yannik Schnitzer, Jasper Van Beers, Coen de Visser, Alessandro Abate

Recent approaches to leveraging deep learning for computing reachable sets of continuous-time dynamical systems have gained popularity over traditional level-set methods, as they overcome the curse of dimensionality. However, as with level-set methods, considerable care needs to be taken in limiting approximation errors, particularly since no guarantees are provided during training on the accuracy of the learned reachable set. To address this limitation, we introduce an epsilon-approximate Hamilton-Jacobi Partial Differential Equation (HJ-PDE), which establishes a relationship between training loss and accuracy of the true reachable set. To formally certify this approximation, we leverage Satisfiability Modulo Theories (SMT) solvers to bound the residual error of the HJ-based loss function across the domain of interest. Leveraging Counter Example Guided Inductive Synthesis (CEGIS), we close the loop around learning and verification, by fine-tuning the neural network on counterexamples found by the SMT solver, thus improving the accuracy of the learned reachable set. To the best of our knowledge, Certified Approximate Reachability (CARe) is the first approach to provide soundness guarantees on learned reachable sets of continuous dynamical systems.

☆ Feature learning from non-Gaussian inputs: the case of Independent Component Analysis in high dimensions

Fabiola Ricci, Lorenzo Bardone, Sebastian Goldt

Deep neural networks learn structured features from complex, non-Gaussian inputs, but the mechanisms behind this process remain poorly understood. Our work is motivated by the observation that the first-layer filters learnt by deep convolutional neural networks from natural images resemble those learnt by independent component analysis (ICA), a simple unsupervised method that seeks the most non-Gaussian projections of its inputs. This similarity suggests that ICA provides a simple, yet principled model for studying feature learning. Here, we leverage this connection to investigate the interplay between data structure and optimisation in feature learning for the most popular ICA algorithm, FastICA, and stochastic gradient descent (SGD), which is used to train deep networks. We rigorously establish that FastICA requires at least $n\gtrsim d^4$ samples to recover a single non-Gaussian direction from $d$-dimensional inputs on a simple synthetic data model. We show that vanilla online SGD outperforms FastICA, and prove that the optimal sample complexity $n \gtrsim d^2$ can be reached by smoothing the loss, albeit in a data-dependent way. We finally demonstrate the existence of a search phase for FastICA on ImageNet, and discuss how the strong non-Gaussianity of said images compensates for the poor sample complexity of FastICA.

☆ DiffScale: Continuous Downscaling and Bias Correction of Subseasonal Wind Speed Forecasts using Diffusion Models

Maximilian Springenberg, Noelia Otero, Yuxin Xue, Jackie Ma

Renewable resources are strongly dependent on local and large-scale weather situations. Skillful subseasonal to seasonal (S2S) forecasts -- beyond two weeks and up to two months -- can offer significant socioeconomic advantages to the energy sector. This study aims to enhance wind speed predictions using a diffusion model with classifier-free guidance to downscale S2S forecasts of surface wind speed. We propose DiffScale, a diffusion model that super-resolves spatial information for continuous downscaling factors and lead times. Leveraging weather priors as guidance for the generative process of diffusion models, we adopt the perspective of conditional probabilities on sampling super-resolved S2S forecasts. We aim to directly estimate the density associated with the target S2S forecasts at different spatial resolutions and lead times without auto-regression or sequence prediction, resulting in an efficient and flexible model. Synthetic experiments were designed to super-resolve wind speed S2S forecasts from the European Center for Medium-Range Weather Forecast (ECMWF) from a coarse resolution to a finer resolution of ERA5 reanalysis data, which serves as a high-resolution target. The innovative aspect of DiffScale lies in its flexibility to downscale arbitrary scaling factors, enabling it to generalize across various grid resolutions and lead times -without retraining the model- while correcting model errors, making it a versatile tool for improving S2S wind speed forecasts. We achieve a significant improvement in prediction quality, outperforming baselines up to week 3.

comment: 28 pages, 18 figures, preprint under review

☆ An End-to-End Comprehensive Gear Fault Diagnosis Method Based on Multi-Scale Feature-Level Fusion Strategy

Bowei Qiao, Hongwei Wang

To satisfy the requirements of the end-to-end fault diagnosis of gears, an integrated intelligent method of fault diagnosis for gears using acceleration signals was proposed, which was based on Gabor-based Adaptive Short-Time Fourier Transform (Gabor-ASTFT) and Dual-Tree Complex Wavelet Transform(DTCWT) algorithms, Dilated Residual structure and feature fusion layer, is proposed in this paper. Initially, the raw one-dimensional acceleration signals collected from the gearbox base using vibration sensors undergo pre-segmentation processing. The Gabor-ASTFT and DTCWT are then applied to convert the original one-dimensional time-domain signals into two-dimensional time-frequency representations, facilitating the preliminary extraction of fault features and obtaining weak feature maps.Subsequently, a dual-channel structure is established using deconvolution and dilated convolution to perform upsampling and downsampling on the feature maps, adjusting their sizes accordingly. A feature fusion layer is then constructed to integrate the dual-channel features, enabling multi-scale analysis of the extracted fault features.Finally, a convolutional neural network (CNN) model incorporating a residual structure is developed to conduct deep feature extraction from the fused feature maps. The extracted features are subsequently fed into a Global Average Pooling(GAP) and a classification function for fault classification. Conducting comparative experiments on different datasets, the proposed method is demonstrated to effectively meet the requirements of end-to-end fault diagnosis for gears.

☆ ZeroMimic: Distilling Robotic Manipulation Skills from Web Videos ICRA 2025

Junyao Shi, Zhuolun Zhao, Tianyou Wang, Ian Pedroza, Amy Luo, Jie Wang, Jason Ma, Dinesh Jayaraman

Many recent advances in robotic manipulation have come through imitation learning, yet these rely largely on mimicking a particularly hard-to-acquire form of demonstrations: those collected on the same robot in the same room with the same objects as the trained policy must handle at test time. In contrast, large pre-recorded human video datasets demonstrating manipulation skills in-the-wild already exist, which contain valuable information for robots. Is it possible to distill a repository of useful robotic skill policies out of such data without any additional requirements on robot-specific demonstrations or exploration? We present the first such system ZeroMimic, that generates immediately deployable image goal-conditioned skill policies for several common categories of manipulation tasks (opening, closing, pouring, pick&place, cutting, and stirring) each capable of acting upon diverse objects and across diverse unseen task setups. ZeroMimic is carefully designed to exploit recent advances in semantic and geometric visual understanding of human videos, together with modern grasp affordance detectors and imitation policy classes. After training ZeroMimic on the popular EpicKitchens dataset of ego-centric human videos, we evaluate its out-of-the-box performance in varied real-world and simulated kitchen settings with two different robot embodiments, demonstrating its impressive abilities to handle these varied tasks. To enable plug-and-play reuse of ZeroMimic policies on other task setups and robots, we release software and policy checkpoints of our skill policies.

comment: ICRA 2025. Project website: https://zeromimic.github.io/

☆ Communication-Efficient and Personalized Federated Foundation Model Fine-Tuning via Tri-Matrix Adaptation

Yongle Li, Bo Liu, Sheng Huang, ZHeng ZHang, Xiaotong Yuan, Richang Hong

In federated learning, fine-tuning pre-trained foundation models poses significant challenges, particularly regarding high communication cost and suboptimal model performance due to data heterogeneity between the clients. To address these issues, this paper introduces communication-efficient federated LoRA adaption (CE-LoRA), a method that employs a tri-factorization low-rank adaptation approach with personalized model parameter aggregation. We first presents a novel LoRA parameter factorization by introducing a small-size dense matrix, which can significantly reduce the communication cost and achieve comparable empirical performance than transferring the low-rank parameter matrix used by existing methods. Without violating data privacy, the server considers the client similarity in both training dataset and model parameter space, and learns personalized weights for model aggregation. Our experiments on various LLM and VLM fine-tuning tasks demonstrate that CE-LoRA not only significantly reduces communication overhead but also improves performance under not independently and identically distributed data conditions. In addition, CE-LoRA improves data privacy protection, effectively mitigating gradient-based data reconstruction attacks.

☆ A Channel-Triggered Backdoor Attack on Wireless Semantic Image Reconstruction

Jialin Wan, Nan Cheng, Jinglong Shen

Despite the transformative impact of deep learning (DL) on wireless communication systems through data-driven end-to-end (E2E) learning, the security vulnerabilities of these systems have been largely overlooked. Unlike the extensively studied image domain, limited research has explored the threat of backdoor attacks on the reconstruction of symbols in semantic communication (SemCom) systems. Previous work has investigated such backdoor attacks at the input level, but these approaches are infeasible in applications with strict input control. In this paper, we propose a novel attack paradigm, termed Channel-Triggered Backdoor Attack (CT-BA), where the backdoor trigger is a specific wireless channel. This attack leverages fundamental physical layer characteristics, making it more covert and potentially more threatening compared to previous input-level attacks. Specifically, we utilize channel gain with different fading distributions or channel noise with different power spectral densities as potential triggers. This approach establishes unprecedented attack flexibility as the adversary can select backdoor triggers from both fading characteristics and noise variations in diverse channel environments. Moreover, during the testing phase, CT-BA enables automatic trigger activation through natural channel variations without requiring active adversary participation. We evaluate the robustness of CT-BA on a ViT-based Joint Source-Channel Coding (JSCC) model across three datasets: MNIST, CIFAR-10, and ImageNet. Furthermore, we apply CT-BA to three typical E2E SemCom systems: BDJSCC, ADJSCC, and JSCCOFDM. Experimental results demonstrate that our attack achieves near-perfect attack success rate (ASR) while maintaining effective stealth. Finally, we discuss potential defense mechanisms against such attacks.

☆ An extrapolated and provably convergent algorithm for nonlinear matrix decomposition with the ReLU function

Nicolas Gillis, Margherita Porcelli, Giovanni Seraghiti

Nonlinear matrix decomposition (NMD) with the ReLU function, denoted ReLU-NMD, is the following problem: given a sparse, nonnegative matrix $X$ and a factorization rank $r$, identify a rank-$r$ matrix $\Theta$ such that $X\approx \max(0,\Theta)$. This decomposition finds application in data compression, matrix completion with entries missing not at random, and manifold learning. The standard ReLU-NMD model minimizes the least squares error, that is, $\|X - \max(0,\Theta)\|_F^2$. The corresponding optimization problem is nondifferentiable and highly nonconvex. This motivated Saul to propose an alternative model, Latent-ReLU-NMD, where a latent variable $Z$ is introduced and satisfies $\max(0,Z)=X$ while minimizing $\|Z - \Theta\|_F^2$ (``A nonlinear matrix decomposition for mining the zeros of sparse data'', SIAM J. Math. Data Sci., 2022). Our first contribution is to show that the two formulations may yield different low-rank solutions $\Theta$; in particular, we show that Latent-ReLU-NMD can be ill-posed when ReLU-NMD is not, meaning that there are instances in which the infimum of Latent-ReLU-NMD is not attained while that of ReLU-NMD is. We also consider another alternative model, called 3B-ReLU-NMD, which parameterizes $\Theta=WH$, where $W$ has $r$ columns and $H$ has $r$ rows, allowing one to get rid of the rank constraint in Latent-ReLU-NMD. Our second contribution is to prove the convergence of a block coordinate descent (BCD) applied to 3B-ReLU-NMD and referred to as BCD-NMD. Our third contribution is a novel extrapolated variant of BCD-NMD, dubbed eBCD-NMD, which we prove is also convergent under mild assumptions. We illustrate the significant acceleration effect of eBCD-NMD compared to BCD-NMD, and also show that eBCD-NMD performs well against the state of the art on synthetic and real-world data sets.

comment: 27 pages. Codes and data available from https://github.com/giovanniseraghiti/ReLU-NMD

☆ Node Embeddings via Neighbor Embeddings

Jan Niklas Böhm, Marius Keute, Alica Guzmán, Sebastian Damrich, Andrew Draganov, Dmitry Kobak

Graph layouts and node embeddings are two distinct paradigms for non-parametric graph representation learning. In the former, nodes are embedded into 2D space for visualization purposes. In the latter, nodes are embedded into a high-dimensional vector space for downstream processing. State-of-the-art algorithms for these two paradigms, force-directed layouts and random-walk-based contrastive learning (such as DeepWalk and node2vec), have little in common. In this work, we show that both paradigms can be approached with a single coherent framework based on established neighbor embedding methods. Specifically, we introduce graph t-SNE, a neighbor embedding method for two-dimensional graph layouts, and graph CNE, a contrastive neighbor embedding method that produces high-dimensional node representations by optimizing the InfoNCE objective. We show that both graph t-SNE and graph CNE strongly outperform state-of-the-art algorithms in terms of local structure preservation, while being conceptually simpler.

☆ When Counterfactual Reasoning Fails: Chaos and Real-World Complexity

Yahya Aalaila, Gerrit Großmann, Sumantrak Mukherjee, Jonas Wahl, Sebastian Vollmer

Counterfactual reasoning, a cornerstone of human cognition and decision-making, is often seen as the 'holy grail' of causal learning, with applications ranging from interpreting machine learning models to promoting algorithmic fairness. While counterfactual reasoning has been extensively studied in contexts where the underlying causal model is well-defined, real-world causal modeling is often hindered by model and parameter uncertainty, observational noise, and chaotic behavior. The reliability of counterfactual analysis in such settings remains largely unexplored. In this work, we investigate the limitations of counterfactual reasoning within the framework of Structural Causal Models. Specifically, we empirically investigate \emph{counterfactual sequence estimation} and highlight cases where it becomes increasingly unreliable. We find that realistic assumptions, such as low degrees of model uncertainty or chaotic dynamics, can result in counterintuitive outcomes, including dramatic deviations between predicted and true counterfactual trajectories. This work urges caution when applying counterfactual reasoning in settings characterized by chaos and uncertainty. Furthermore, it raises the question of whether certain systems may pose fundamental limitations on the ability to answer counterfactual questions about their behavior.

☆ Conformal uncertainty quantification to evaluate predictive fairness of foundation AI model for skin lesion classes across patient demographics

Swarnava Bhattacharyya, Umapada Pal, Tapabrata Chakraborti

Deep learning based diagnostic AI systems based on medical images are starting to provide similar performance as human experts. However these data hungry complex systems are inherently black boxes and therefore slow to be adopted for high risk applications like healthcare. This problem of lack of transparency is exacerbated in the case of recent large foundation models, which are trained in a self supervised manner on millions of data points to provide robust generalisation across a range of downstream tasks, but the embeddings generated from them happen through a process that is not interpretable, and hence not easily trustable for clinical applications. To address this timely issue, we deploy conformal analysis to quantify the predictive uncertainty of a vision transformer (ViT) based foundation model across patient demographics with respect to sex, age and ethnicity for the tasks of skin lesion classification using several public benchmark datasets. The significant advantage of this method is that conformal analysis is method independent and it not only provides a coverage guarantee at population level but also provides an uncertainty score for each individual. We used a model-agnostic dynamic F1-score-based sampling during model training, which helped to stabilize the class imbalance and we investigate the effects on uncertainty quantification (UQ) with or without this bias mitigation step. Thus we show how this can be used as a fairness metric to evaluate the robustness of the feature embeddings of the foundation model (Google DermFoundation) and thus advance the trustworthiness and fairness of clinical AI.

☆ Free Parametrization of L2-bounded State Space Models

Leonardo Massai, Giancarlo Ferrari-Trecate

Structured state-space models (SSMs) have emerged as a powerful architecture in machine learning and control, featuring stacked layers where each consists of a linear time-invariant (LTI) discrete-time system followed by a nonlinearity. While SSMs offer computational efficiency and excel in long-sequence predictions, their widespread adoption in applications like system identification and optimal control is hindered by the challenge of ensuring their stability and robustness properties. We introduce L2RU, a novel parametrization of SSMs that guarantees input-output stability and robustness by enforcing a prescribed L-bound for all parameter values. This design eliminates the need for complex constraints, allowing unconstrained optimization over L2RUs by using standard methods such as gradient descent. Leveraging tools from system theory and convex optimization, we derive a non-conservative parametrization of square discrete-time LTI systems with a specified L2-bound, forming the foundation of the L2RU architecture. Additionally, we enhance its performance with a bespoke initialization strategy optimized for long input sequences. Through a system identification task, we validate L2RU's superior performance, showcasing its potential in learning and control applications.

comment: 8 pages

☆ An extension of linear self-attention for in-context learning

Katsuyuki Hagiwara

In-context learning is a remarkable property of transformers and has been the focus of recent research. An attention mechanism is a key component in transformers, in which an attention matrix encodes relationships between words in a sentence and is used as weights for words in a sentence. This mechanism is effective for capturing language representations. However, it is questionable whether naive self-attention is suitable for in-context learning in general tasks, since the computation implemented by self-attention is somewhat restrictive in terms of matrix multiplication. In fact, we may need appropriate input form designs when considering heuristic implementations of computational algorithms. In this paper, in case of linear self-attention, we extend it by introducing a bias matrix in addition to a weight matrix for an input. Despite the simple extension, the extended linear self-attention can output any constant matrix, input matrix and multiplications of two or three matrices in the input. Note that the second property implies that it can be a skip connection. Therefore, flexible matrix manipulations can be implemented by connecting the extended linear self-attention components. As an example of implementation using the extended linear self-attention, we show a heuristic construction of a batch-type gradient descent of ridge regression under a reasonable input form.

☆ Adaptive Attention-Based Model for 5G Radio-based Outdoor Localization

Ilayda Yaman, Guoda Tian, Fredrik Tufvesson, Ove Edfors, Zhengya Zhang, Liang Liu

Radio-based localization in dynamic environments, such as urban and vehicular settings, requires systems that can efficiently adapt to varying signal conditions and environmental changes. Factors such as multipath interference and obstructions introduce different levels of complexity that affect the accuracy of the localization. Although generalized models offer broad applicability, they often struggle to capture the nuances of specific environments, leading to suboptimal performance in real-world deployments. In contrast, specialized models can be tailored to particular conditions, enabling more precise localization by effectively handling domain-specific variations and noise patterns. However, deploying multiple specialized models requires an efficient mechanism to select the most appropriate one for a given scenario. In this work, we develop an adaptive localization framework that combines shallow attention-based models with a router/switching mechanism based on a single-layer perceptron (SLP). This enables seamless transitions between specialized localization models optimized for different conditions, balancing accuracy, computational efficiency, and robustness to environmental variations. We design three low-complex localization models tailored for distinct scenarios, optimized for reduced computational complexity, test time, and model size. The router dynamically selects the most suitable model based on real-time input characteristics. The proposed framework is validated using real-world vehicle localization data collected from a massive MIMO base station (BS), demonstrating its ability to seamlessly adapt to diverse deployment conditions while maintaining high localization accuracy.

comment: 6 pages, 6 figures

☆ Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks

Fabian L. Thiemann, Thiago Reschützegger, Massimiliano Esposito, Tseden Taddese, Juan D. Olarte-Plata, Fausto Martelli

Molecular dynamics (MD) simulations play a crucial role in scientific research. Yet their computational cost often limits the timescales and system sizes that can be explored. Most data-driven efforts have been focused on reducing the computational cost of accurate interatomic forces required for solving the equations of motion. Despite their success, however, these machine learning interatomic potentials (MLIPs) are still bound to small time-steps. In this work, we introduce TrajCast, a transferable and data-efficient framework based on autoregressive equivariant message passing networks that directly updates atomic positions and velocities lifting the constraints imposed by traditional numerical integration. We benchmark our framework across various systems, including a small molecule, crystalline material, and bulk liquid, demonstrating excellent agreement with reference MD simulations for structural, dynamical, and energetic properties. Depending on the system, TrajCast allows for forecast intervals up to $30\times$ larger than traditional MD time-steps, generating over 15 ns of trajectory data per day for a solid with more than 4,000 atoms. By enabling efficient large-scale simulations over extended timescales, TrajCast can accelerate materials discovery and explore physical phenomena beyond the reach of traditional simulations and experiments. An open-source implementation of TrajCast is accessible under https://github.com/IBM/trajcast.

comment: 25 pages total (19 manuscript, 6 SI). 5 figures in manuscript, 3 figures and 2 tables in SI

☆ Evaluation of (Un-)Supervised Machine Learning Methods for GNSS Interference Classification with Real-World Data Discrepancies

Lucas Heublein, Nisha L. Raichur, Tobias Feigl, Tobias Brieger, Fin Heuer, Lennart Asbach, Alexander Rügamer, Felix Ott

The accuracy and reliability of vehicle localization on roads are crucial for applications such as self-driving cars, toll systems, and digital tachographs. To achieve accurate positioning, vehicles typically use global navigation satellite system (GNSS) receivers to validate their absolute positions. However, GNSS-based positioning can be compromised by interference signals, necessitating the identification, classification, determination of purpose, and localization of such interference to mitigate or eliminate it. Recent approaches based on machine learning (ML) have shown superior performance in monitoring interference. However, their feasibility in real-world applications and environments has yet to be assessed. Effective implementation of ML techniques requires training datasets that incorporate realistic interference signals, including real-world noise and potential multipath effects that may occur between transmitter, receiver, and satellite in the operational area. Additionally, these datasets require reference labels. Creating such datasets is often challenging due to legal restrictions, as causing interference to GNSS sources is strictly prohibited. Consequently, the performance of ML-based methods in practical applications remains unclear. To address this gap, we describe a series of large-scale measurement campaigns conducted in real-world settings at two highway locations in Germany and the Seetal Alps in Austria, and in large-scale controlled indoor environments. We evaluate the latest supervised ML-based methods to report on their performance in real-world settings and present the applicability of pseudo-labeling for unsupervised learning. We demonstrate the challenges of combining datasets due to data discrepancies and evaluate outlier detection, domain adaptation, and data augmentation techniques to present the models' capabilities to adapt to changes in the datasets.

comment: 34 pages, 25 figures

☆ Accelerating High-Efficiency Organic Photovoltaic Discovery via Pretrained Graph Neural Networks and Generative Reinforcement Learning ICLR 2025

Jiangjie Qiu, Hou Hei Lam, Xiuyuan Hu, Wentao Li, Siwei Fu, Fankun Zeng, Hao Zhang, Xiaonan Wang

Organic photovoltaic (OPV) materials offer a promising avenue toward cost-effective solar energy utilization. However, optimizing donor-acceptor (D-A) combinations to achieve high power conversion efficiency (PCE) remains a significant challenge. In this work, we propose a framework that integrates large-scale pretraining of graph neural networks (GNNs) with a GPT-2 (Generative Pretrained Transformer 2)-based reinforcement learning (RL) strategy to design OPV molecules with potentially high PCE. This approach produces candidate molecules with predicted efficiencies approaching 21\%, although further experimental validation is required. Moreover, we conducted a preliminary fragment-level analysis to identify structural motifs recognized by the RL model that may contribute to enhanced PCE, thus providing design guidelines for the broader research community. To facilitate continued discovery, we are building the largest open-source OPV dataset to date, expected to include nearly 3,000 donor-acceptor pairs. Finally, we discuss plans to collaborate with experimental teams on synthesizing and characterizing AI-designed molecules, which will provide new data to refine and improve our predictive and generative models.

comment: AI for Accelerated Materials Design - ICLR 2025

☆ Time-Series Forecasting via Topological Information Supervised Framework with Efficient Topological Feature Learning

ZiXin Lin, Nur Fariha Syaqina Zulkepli

Topological Data Analysis (TDA) has emerged as a powerful tool for extracting meaningful features from complex data structures, driving significant advancements in fields such as neuroscience, biology, machine learning, and financial modeling. Despite its success, the integration of TDA with time-series prediction remains underexplored due to three primary challenges: the limited utilization of temporal dependencies within topological features, computational bottlenecks associated with persistent homology, and the deterministic nature of TDA pipelines restricting generalized feature learning. This study addresses these challenges by proposing the Topological Information Supervised (TIS) Prediction framework, which leverages neural networks and Conditional Generative Adversarial Networks (CGANs) to generate synthetic topological features, preserving their distribution while significantly reducing computational time. We propose a novel training strategy that integrates topological consistency loss to improve the predictive accuracy of deep learning models. Specifically, we introduce two state-of-the-art models, TIS-BiGRU and TIS-Informer, designed to capture short-term and long-term temporal dependencies, respectively. Comparative experimental results demonstrate the superior performance of TIS models over conventional predictors, validating the effectiveness of integrating topological information. This work not only advances TDA-based time-series prediction but also opens new avenues for utilizing topological features in deep learning architectures.

☆ THEMIS: Towards Practical Intellectual Property Protection for Post-Deployment On-Device Deep Learning Models USENIX Security

Yujin Huang, Zhi Zhang, Qingchuan Zhao, Xingliang Yuan, Chunyang Chen

On-device deep learning (DL) has rapidly gained adoption in mobile apps, offering the benefits of offline model inference and user privacy preservation over cloud-based approaches. However, it inevitably stores models on user devices, introducing new vulnerabilities, particularly model-stealing attacks and intellectual property infringement. While system-level protections like Trusted Execution Environments (TEEs) provide a robust solution, practical challenges remain in achieving scalable on-device DL model protection, including complexities in supporting third-party models and limited adoption in current mobile solutions. Advancements in TEE-enabled hardware, such as NVIDIA's GPU-based TEEs, may address these obstacles in the future. Currently, watermarking serves as a common defense against model theft but also faces challenges here as many mobile app developers lack corresponding machine learning expertise and the inherent read-only and inference-only nature of on-device DL models prevents third parties like app stores from implementing existing watermarking techniques in post-deployment models. To protect the intellectual property of on-device DL models, in this paper, we propose THEMIS, an automatic tool that lifts the read-only restriction of on-device DL models by reconstructing their writable counterparts and leverages the untrainable nature of on-device DL models to solve watermark parameters and protect the model owner's intellectual property. Extensive experimental results across various datasets and model structures show the superiority of THEMIS in terms of different metrics. Further, an empirical investigation of 403 real-world DL mobile apps from Google Play is performed with a success rate of 81.14%, showing the practicality of THEMIS.

comment: To Appear in the 34th USENIX Security Symposium, August 13-15, 2025

☆ Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model

Dizhan Xue, Jing Cui, Shengsheng Qian, Chuanrui Hu, Changsheng Xu

Short-video platforms have gained immense popularity, captivating the interest of millions, if not billions, of users globally. Recently, researchers have highlighted the significance of analyzing the propagation of short-videos, which typically involves discovering commercial values, public opinions, user behaviors, etc. This paper proposes a new Short-video Propagation Influence Rating (SPIR) task and aims to promote SPIR from both the dataset and method perspectives. First, we propose a new Cross-platform Short-Video (XS-Video) dataset, which aims to provide a large-scale and real-world short-video propagation network across various platforms to facilitate the research on short-video propagation. Our XS-Video dataset includes 117,720 videos, 381,926 samples, and 535 topics across 5 biggest Chinese platforms, annotated with the propagation influence from level 0 to 9. To the best of our knowledge, this is the first large-scale short-video dataset that contains cross-platform data or provides all of the views, likes, shares, collects, fans, comments, and comment content. Second, we propose a Large Graph Model (LGM) named NetGPT, based on a novel three-stage training mechanism, to bridge heterogeneous graph-structured data with the powerful reasoning ability and knowledge of Large Language Models (LLMs). Our NetGPT can comprehend and analyze the short-video propagation graph, enabling it to predict the long-term propagation influence of short-videos. Comprehensive experimental results evaluated by both classification and regression metrics on our XS-Video dataset indicate the superiority of our method for SPIR.

☆ Integral regularization PINNs for evolution equations

Xiaodong Feng, Haojiong Shangguan, Tao Tang, Xiaoliang Wan

Evolution equations, including both ordinary differential equations (ODEs) and partial differential equations (PDEs), play a pivotal role in modeling dynamic systems. However, achieving accurate long-time integration for these equations remains a significant challenge. While physics-informed neural networks (PINNs) provide a mesh-free framework for solving PDEs, they often suffer from temporal error accumulation, which limits their effectiveness in capturing long-time behaviors. To alleviate this issue, we propose integral regularization PINNs (IR-PINNs), a novel approach that enhances temporal accuracy by incorporating an integral-based residual term into the loss function. This method divides the entire time interval into smaller sub-intervals and enforces constraints over these sub-intervals, thereby improving the resolution and correlation of temporal dynamics. Furthermore, IR-PINNs leverage adaptive sampling to dynamically refine the distribution of collocation points based on the evolving solution, ensuring higher accuracy in regions with sharp gradients or rapid variations. Numerical experiments on benchmark problems demonstrate that IR-PINNs outperform original PINNs and other state-of-the-art methods in capturing long-time behaviors, offering a robust and accurate solution for evolution equations.

☆ PDSL: Privacy-Preserved Decentralized Stochastic Learning with Heterogeneous Data Distribution

Lina Wang, Yunsheng Yuan, Chunxiao Wang, Feng Li

In the paradigm of decentralized learning, a group of agents collaborates to learn a global model using distributed datasets without a central server. However, due to the heterogeneity of the local data across the different agents, learning a robust global model is rather challenging. Moreover, the collaboration of the agents relies on their gradient information exchange, which poses a risk of privacy leakage. In this paper, to address these issues, we propose PDSL, a novel privacy-preserved decentralized stochastic learning algorithm with heterogeneous data distribution. On one hand, we innovate in utilizing the notion of Shapley values such that each agent can precisely measure the contributions of its heterogeneous neighbors to the global learning goal; on the other hand, we leverage the notion of differential privacy to prevent each agent from suffering privacy leakage when it contributes gradient information to its neighbors. We conduct both solid theoretical analysis and extensive experiments to demonstrate the efficacy of our PDSL algorithm in terms of privacy preservation and convergence.

☆ Unimodal-driven Distillation in Multimodal Emotion Recognition with Dynamic Fusion

Jiagen Li, Rui Yu, Huihao Huang, Huaicheng Yan

Multimodal Emotion Recognition in Conversations (MERC) identifies emotional states across text, audio and video, which is essential for intelligent dialogue systems and opinion analysis. Existing methods emphasize heterogeneous modal fusion directly for cross-modal integration, but often suffer from disorientation in multimodal learning due to modal heterogeneity and lack of instructive guidance. In this work, we propose SUMMER, a novel heterogeneous multimodal integration framework leveraging Mixture of Experts with Hierarchical Cross-modal Fusion and Interactive Knowledge Distillation. Key components include a Sparse Dynamic Mixture of Experts (SDMoE) for capturing dynamic token-wise interactions, a Hierarchical Cross-Modal Fusion (HCMF) for effective fusion of heterogeneous modalities, and Interactive Knowledge Distillation (IKD), which uses a pre-trained unimodal teacher to guide multimodal fusion in latent and logit spaces. Experiments on IEMOCAP and MELD show SUMMER outperforms state-of-the-art methods, particularly in recognizing minority and semantically similar emotions.

☆ Steering Large Agent Populations using Mean-Field Schrodinger Bridges with Gaussian Mixture Models

George Rapakoulias, Ali Reza Pedram, Panagiotis Tsiotras

The Mean-Field Schrodinger Bridge (MFSB) problem is an optimization problem aiming to find the minimum effort control policy to drive a McKean-Vlassov stochastic differential equation from one probability measure to another. In the context of multiagent control, the objective is to control the configuration of a swarm of identical, interacting cooperative agents, as captured by the time-varying probability measure of their state. Available methods for solving this problem for distributions with continuous support rely either on spatial discretizations of the problem's domain or on approximating optimal solutions using neural networks trained through stochastic optimization schemes. For agents following Linear Time-Varying dynamics, and for Gaussian Mixture Model boundary distributions, we propose a highly efficient parameterization to approximate the solutions of the corresponding MFSB in closed form, without any learning steps. Our proposed approach consists of a mixture of elementary policies, each solving a Gaussian-to-Gaussian Covariance Steering problem from the components of the initial to the components of the terminal mixture. Leveraging the semidefinite formulation of the Covariance Steering problem, our proposed solver can handle probabilistic hard constraints on the system's state, while maintaining numerical tractability. We illustrate our approach on a variety of numerical examples.

☆ A Low-complexity Structured Neural Network to Realize States of Dynamical Systems

Hansaka Aluvihare, Levi Lingsch, Xianqi Li, Sirani M. Perera

Data-driven learning is rapidly evolving and places a new perspective on realizing state-space dynamical systems. However, dynamical systems derived from nonlinear ordinary differential equations (ODEs) suffer from limitations in computational efficiency. Thus, this paper stems from data-driven learning to advance states of dynamical systems utilizing a structured neural network (StNN). The proposed learning technique also seeks to identify an optimal, low-complexity operator to solve dynamical systems, the so-called Hankel operator, derived from time-delay measurements. Thus, we utilize the StNN based on the Hankel operator to solve dynamical systems as an alternative to existing data-driven techniques. We show that the proposed StNN reduces the number of parameters and computational complexity compared with the conventional neural networks and also with the classical data-driven techniques, such as Sparse Identification of Nonlinear Dynamics (SINDy) and Hankel Alternative view of Koopman (HAVOK), which is commonly known as delay-Dynamic Mode Decomposition(DMD) or Hankel-DMD. More specifically, we present numerical simulations to solve dynamical systems utilizing the StNN based on the Hankel operator beginning from the fundamental Lotka-Volterra model, where we compare the StNN with the LEarning Across Dynamical Systems (LEADS), and extend our analysis to highly nonlinear and chaotic Lorenz systems, comparing the StNN with conventional neural networks, SINDy, and HAVOK. Hence, we show that the proposed StNN paves the way for realizing state-space dynamical systems with a low-complexity learning algorithm, enabling prediction and understanding of future states.

comment: 20 pages, 6 figures

☆ MKA: Leveraging Cross-Lingual Consensus for Model Abstention ICLR 2025

Sharad Duwal

Reliability of LLMs is questionable even as they get better at more tasks. A wider adoption of LLMs is contingent on whether they are usably factual. And if they are not, on whether they can properly calibrate their confidence in their responses. This work focuses on utilizing the multilingual knowledge of an LLM to inform its decision to abstain or answer when prompted. We develop a multilingual pipeline to calibrate the model's confidence and let it abstain when uncertain. We run several multilingual models through the pipeline to profile them across different languages. We find that the performance of the pipeline varies by model and language, but that in general they benefit from it. This is evidenced by the accuracy improvement of $71.2\%$ for Bengali over a baseline performance without the pipeline. Even a high-resource language like English sees a $15.5\%$ improvement. These results hint at possible further improvements.

comment: To appear in Building Trust Workshop at ICLR 2025

☆ Data-Driven Forecasting of High-Dimensional Transient and Stationary Processes via Space-Time Projection

Oliver T. Schmidt

Space-Time Projection (STP) is introduced as a data-driven forecasting approach for high-dimensional and time-resolved data. The method computes extended space-time proper orthogonal modes from training data spanning a prediction horizon comprising both hindcast and forecast intervals. Forecasts are then generated by projecting the hindcast portion of these modes onto new data, simultaneously leveraging their orthogonality and optimal correlation with the forecast extension. Rooted in Proper Orthogonal Decomposition (POD) theory, dimensionality reduction and time-delay embedding are intrinsic to the approach. For a given ensemble and fixed prediction horizon, the only tunable parameter is the truncation rank--no additional hyperparameters are required. The hindcast accuracy serves as a reliable indicator for short-term forecast accuracy and establishes a lower bound on forecast errors. The efficacy of the method is demonstrated using two datasets: transient, highly anisotropic simulations of supernova explosions in a turbulent interstellar medium, and experimental velocity fields of a turbulent high-subsonic engineering flow. In a comparative study with standard Long Short-Term Memory (LSTM) neural networks--acknowledging that alternative architectures or training strategies may yield different outcomes--the method consistently provided more accurate forecasts. Considering its simplicity and robust performance, STP offers an interpretable and competitive benchmark for forecasting high-dimensional transient and chaotic processes, relying purely on spatiotemporal correlation information.

☆ Dynamic Operating System Scheduling Using Double DQN: A Reinforcement Learning Approach to Task Optimization

Xiaoxuan Sun, Yifei Duan, Yingnan Deng, Fan Guo, Guohui Cai, Yuting Peng

In this paper, an operating system scheduling algorithm based on Double DQN (Double Deep Q network) is proposed, and its performance under different task types and system loads is verified by experiments. Compared with the traditional scheduling algorithm, the algorithm based on Double DQN can dynamically adjust the task priority and resource allocation strategy, thus improving the task completion efficiency, system throughput, and response speed. The experimental results show that the Double DQN algorithm has high scheduling performance under light load, medium load and heavy load scenarios, especially when dealing with I/O intensive tasks, and can effectively reduce task completion time and system response time. In addition, the algorithm also shows high optimization ability in resource utilization and can intelligently adjust resource allocation according to the system state, avoiding resource waste and excessive load. Future studies will further explore the application of the algorithm in more complex systems, especially scheduling optimization in cloud computing and large-scale distributed environments, combining factors such as network latency and energy efficiency to improve the overall performance and adaptability of the algorithm.

☆ Scalable Geometric Learning with Correlation-Based Functional Brain Networks

Kisung You, Hae-Jeong Park

The correlation matrix is a central representation of functional brain networks in neuroimaging. Traditional analyses often treat pairwise interactions independently in a Euclidean setting, overlooking the intrinsic geometry of correlation matrices. While earlier attempts have embraced the quotient geometry of the correlation manifold, they remain limited by computational inefficiency and numerical instability, particularly in high-dimensional contexts. This paper presents a novel geometric framework that employs diffeomorphic transformations to embed correlation matrices into a Euclidean space, preserving salient manifold properties and enabling large-scale analyses. The proposed method integrates with established learning algorithms - regression, dimensionality reduction, and clustering - and extends naturally to population-level inference of brain networks. Simulation studies demonstrate both improved computational speed and enhanced accuracy compared to conventional manifold-based approaches. Moreover, applications in real neuroimaging scenarios illustrate the framework's utility, enhancing behavior score prediction, subject fingerprinting in resting-state fMRI, and hypothesis testing in electroencephalogram data. An open-source MATLAB toolbox is provided to facilitate broader adoption and advance the application of correlation geometry in functional brain network research.

☆ A Survey of Reinforcement Learning-Based Motion Planning for Autonomous Driving: Lessons Learned from a Driving Task Perspective

Zhuoren Li, Guizhe Jin, Ran Yu, Zhiwen Chen, Nan Li, Wei Han, Lu Xiong, Bo Leng, Jia Hu, Ilya Kolmanovsky, Dimitar Filev

Reinforcement learning (RL), with its ability to explore and optimize policies in complex, dynamic decision-making tasks, has emerged as a promising approach to addressing motion planning (MoP) challenges in autonomous driving (AD). Despite rapid advancements in RL and AD, a systematic description and interpretation of the RL design process tailored to diverse driving tasks remains underdeveloped. This survey provides a comprehensive review of RL-based MoP for AD, focusing on lessons from task-specific perspectives. We first outline the fundamentals of RL methodologies, and then survey their applications in MoP, analyzing scenario-specific features and task requirements to shed light on their influence on RL design choices. Building on this analysis, we summarize key design experiences, extract insights from various driving task applications, and provide guidance for future implementations. Additionally, we examine the frontier challenges in RL-based MoP, review recent efforts to addresse these challenges, and propose strategies for overcoming unresolved issues.

comment: 21 pages, 5 figures

☆ Learning a Single Index Model from Anisotropic Data with vanilla Stochastic Gradient Descent

Guillaume Braun, Minh Ha Quang, Masaaki Imaizumi

We investigate the problem of learning a Single Index Model (SIM)- a popular model for studying the ability of neural networks to learn features - from anisotropic Gaussian inputs by training a neuron using vanilla Stochastic Gradient Descent (SGD). While the isotropic case has been extensively studied, the anisotropic case has received less attention and the impact of the covariance matrix on the learning dynamics remains unclear. For instance, Mousavi-Hosseini et al. (2023b) proposed a spherical SGD that requires a separate estimation of the data covariance matrix, thereby oversimplifying the influence of covariance. In this study, we analyze the learning dynamics of vanilla SGD under the SIM with anisotropic input data, demonstrating that vanilla SGD automatically adapts to the data's covariance structure. Leveraging these results, we derive upper and lower bounds on the sample complexity using a notion of effective dimension that is determined by the structure of the covariance matrix instead of the input data dimension.

♻ ☆ Evil twins are not that evil: Qualitative insights into machine-generated prompts

Nathanaël Carraz Rakotonirina, Corentin Kervadec, Francesca Franzon, Marco Baroni

It has been widely observed that language models (LMs) respond in predictable ways to algorithmically generated prompts that are seemingly unintelligible. This is both a sign that we lack a full understanding of how LMs work, and a practical challenge, because opaqueness can be exploited for harmful uses of LMs, such as jailbreaking. We present the first thorough analysis of opaque machine-generated prompts, or autoprompts, pertaining to 6 LMs of different sizes and families. We find that machine-generated prompts are characterized by a last token that is often intelligible and strongly affects the generation. A small but consistent proportion of the previous tokens are prunable, probably appearing in the prompt as a by-product of the fact that the optimization process fixes the number of tokens. The remaining tokens fall into two categories: filler tokens, which can be replaced with semantically unrelated substitutes, and keywords, that tend to have at least a loose semantic relation with the generation, although they do not engage in well-formed syntactic relations with it. Additionally, human experts can reliably identify the most influential tokens in an autoprompt a posteriori, suggesting these prompts are not entirely opaque. Finally, some of the ablations we applied to autoprompts yield similar effects in natural language inputs, suggesting that autoprompts emerge naturally from the way LMs process linguistic inputs in general.

♻ ☆ The impact of internal variability on benchmarking deep learning climate emulators

Björn Lütjens, Raffaele Ferrari, Duncan Watson-Parris, Noelle Selin

Full-complexity Earth system models (ESMs) are computationally very expensive, limiting their use in exploring the climate outcomes of multiple emission pathways. More efficient emulators that approximate ESMs can directly map emissions onto climate outcomes, and benchmarks are being used to evaluate their accuracy on standardized tasks and datasets. We investigate a popular benchmark in data-driven climate emulation, ClimateBench, on which deep learning-based emulators are currently achieving the best performance. We compare these deep learning emulators with a linear regression-based emulator, akin to pattern scaling, and show that it outperforms the incumbent 100M-parameter deep learning foundation model, ClimaX, on 3 out of 4 regionally-resolved climate variables, notably surface temperature and precipitation. While emulating surface temperature is expected to be predominantly linear, this result is surprising for emulating precipitation. Precipitation is a much more noisy variable, and we show that deep learning emulators can overfit to internal variability noise at low frequencies, degrading their performance in comparison to a linear emulator. We address the issue of overfitting by increasing the number of climate simulations per emission pathway (from 3 to 50) and updating the benchmark targets with the respective ensemble averages from the MPI-ESM1.2-LR model. Using the new targets, we show that linear pattern scaling continues to be more accurate on temperature, but can be outperformed by a deep learning-based technique for emulating precipitation. We publish our code and data at github.com/blutjens/climate-emulator.

♻ ☆ Inductive Moment Matching

Linqi Zhou, Stefano Ermon, Jiaming Song

Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.

♻ ☆ CryptoPulse: Short-Term Cryptocurrency Forecasting with Dual-Prediction and Cross-Correlated Market Indicators

Amit Kumar, Taoran Ji

Cryptocurrencies fluctuate in markets with high price volatility, posing significant challenges for investors. To aid in informed decision-making, systems predicting cryptocurrency market movements have been developed, typically focusing on historical patterns. However, these methods often overlook three critical factors influencing market dynamics: 1) the macro investing environment, reflected in major cryptocurrency fluctuations affecting collaborative investor behaviors; 2) overall market sentiment, heavily influenced by news impacting investor strategies; and 3) technical indicators, offering insights into overbought or oversold conditions, momentum, and market trends, which are crucial for short-term price movements. This paper proposes a dual prediction mechanism that forecasts the next day's closing price by incorporating macroeconomic fluctuations, technical indicators, and individual cryptocurrency price changes. Additionally, a novel refinement mechanism enhances predictions through market sentiment-based rescaling and fusion. Experiments demonstrate that the proposed model achieves state-of-the-art performance, consistently outperforming ten comparison methods.

comment: 10

♻ ☆ A distance for mixed-variable and hierarchical domains with meta variables

Edward Hallé-Hannan, Charles Audet, Youssef Diouane, Sébastien Le Digabel, Paul Saves

Heterogeneous datasets emerge in various machine learning and optimization applications that feature different input sources, types or formats. Most models or methods do not natively tackle heterogeneity. Hence, such datasets are often partitioned into smaller and simpler ones, which may limit the generalizability or performance, especially when data is limited. The first main contribution of this work is a modeling framework that generalizes hierarchical, tree-structured, variable-size or conditional search frameworks. The framework models mixed-variable and hierarchical domains in which variables may be continuous, integer, or categorical, with some identified as meta when they influence the structure of the problem. The second main contribution is a novel distance that compares any pair of mixed-variable points that do not share the same variables, allowing to use whole heterogeneous datasets that reside in mixed-variable and hierarchical domains with meta variables. The contributions are illustrated through regression and classification experiments using simple distance-based models applied to datasets of hyperparameters with corresponding performance scores.

comment: 29 pages (without references), 12 figures, 5 tables, data and scripts available at https://github.com/bbopt/graph_distance

♻ ☆ Evolutionary Optimization of Physics-Informed Neural Networks: Survey and Prospects

Jian Cheng Wong, Abhishek Gupta, Chin Chun Ooi, Pao-Hsiung Chiu, Jiao Liu, Yew-Soon Ong

Deep learning models trained on finite data lack a complete understanding of the physical world. On the other hand, physics-informed neural networks (PINNs) are infused with such knowledge through the incorporation of mathematically expressible laws of nature into their training loss function. By complying with physical laws, PINNs provide advantages over purely data-driven models in limited-data regimes. This feature has propelled them to the forefront of scientific machine learning, a domain characterized by scarce and costly data. However, the vision of accurate physics-informed learning comes with significant challenges. This review examines PINNs for the first time in terms of model optimization and generalization, shedding light on the need for new algorithmic advances to overcome issues pertaining to the training speed, precision, and generalizability of today's PINN models. Of particular interest are the gradient-free methods of neuroevolution for optimizing the uniquely complex loss landscapes arising in PINN training. Methods synergizing gradient descent and neuroevolution for discovering bespoke neural architectures and balancing multiple conflicting terms in physics-informed learning objectives are positioned as important avenues for future research. Yet another exciting track is to cast neuroevolution as a meta-learner of generalizable PINN models.

comment: 20 pages, 8 figures, 1 table

♻ ☆ DICE: Discrete Inversion Enabling Controllable Editing for Multinomial Diffusion and Masked Generative Models CVPR 2025

Xiaoxiao He, Ligong Han, Quan Dao, Song Wen, Minhao Bai, Di Liu, Han Zhang, Martin Renqiang Min, Felix Juefei-Xu, Chaowei Tan, Bo Liu, Kang Li, Hongdong Li, Junzhou Huang, Faez Ahmed, Akash Srivastava, Dimitris Metaxas

Discrete diffusion models have achieved success in tasks like image generation and masked language modeling but face limitations in controlled content editing. We introduce DICE (Discrete Inversion for Controllable Editing), the first approach to enable precise inversion for discrete diffusion models, including multinomial diffusion and masked generative models. By recording noise sequences and masking patterns during the reverse diffusion process, DICE enables accurate reconstruction and flexible editing of discrete data without the need for predefined masks or attention manipulation. We demonstrate the effectiveness of DICE across both image and text domains, evaluating it on models such as VQ-Diffusion, Paella, and RoBERTa. Our results show that DICE preserves high data fidelity while enhancing editing capabilities, offering new opportunities for fine-grained content manipulation in discrete spaces.

comment: Project webpage: https://hexiaoxiao-cs.github.io/DICE/. This paper was accepted to CVPR 2025 but later desk-rejected post camera-ready, due to a withdrawal from ICLR made 14 days before reviewer assignment

♻ ☆ Cascade Reward Sampling for Efficient Decoding-Time Alignment

Bolian Li, Yifan Wang, Anamika Lochab, Ananth Grama, Ruqi Zhang

Aligning large language models (LLMs) with human preferences is essential for their applications. Recently, decoding-time alignment has emerged as an effective plug-and-play technique that avoids fine-tuning model parameters. This approach retains the general utility of pretrained LLMs but often suffers from significant inefficiencies during decoding, primarily due to wasted token generation and excessive reward evaluations. To address these challenges, we introduce Cascade Reward Sampling (CARDS) to resolve both efficiency bottlenecks in decoding-time alignment. Specifically, we develop a segment-level rejection sampling algorithm that minimizes redundant computations of both LLMs and reward models (RMs). Central to CARDS is an uncertainty-based segmentation mechanism, which ensures the accuracy of RMs evaluations on incomplete segments. Furthermore, we provide a detailed analysis of reward scores on segments to elucidate the improved alignment performance. Experimental results demonstrate that CARDS significantly improves decoding efficiency, alignment quality, and general utility compared to existing decoding-time alignment methods, achieving approximately a 70% reduction in decoding time and over 90% win-ties in utility and safety benchmarks.

♻ ☆ Distributed Fractional Bayesian Learning for Adaptive Optimization

Yaqun Yang, Jinlong Lei, Guanghui Wen, Yiguang Hong

This paper considers a distributed adaptive optimization problem, where all agents only have access to their local cost functions with a common unknown parameter, whereas they mean to collaboratively estimate the true parameter and find the optimal solution over a connected network. A general mathematical framework for such a problem has not been studied yet. We aim to provide valuable insights for addressing parameter uncertainty in distributed optimization problems and simultaneously find the optimal solution. Thus, we propose a novel Prediction while Optimization scheme, which utilizes distributed fractional Bayesian learning through weighted averaging on the log-beliefs to update the beliefs of unknown parameters, and distributed gradient descent for renewing the estimation of the optimal solution. Then under suitable assumptions, we prove that all agents' beliefs and decision variables converge almost surely to the true parameter and the optimal solution under the true parameter, respectively. We further establish a sublinear convergence rate for the belief sequence. Finally, numerical experiments are implemented to corroborate the theoretical analysis.

♻ ☆ Studying the Interplay Between the Actor and Critic Representations in Reinforcement Learning ICLR 2025

Samuel Garcin, Trevor McInroe, Pablo Samuel Castro, Prakash Panangaden, Christopher G. Lucas, David Abel, Stefano V. Albrecht

Extracting relevant information from a stream of high-dimensional observations is a central challenge for deep reinforcement learning agents. Actor-critic algorithms add further complexity to this challenge, as it is often unclear whether the same information will be relevant to both the actor and the critic. To this end, we here explore the principles that underlie effective representations for the actor and for the critic in on-policy algorithms. We focus our study on understanding whether the actor and critic will benefit from separate, rather than shared, representations. Our primary finding is that when separated, the representations for the actor and critic systematically specialise in extracting different types of information from the environment -- the actor's representation tends to focus on action-relevant information, while the critic's representation specialises in encoding value and dynamics information. We conduct a rigourous empirical study to understand how different representation learning approaches affect the actor and critic's specialisations and their downstream performance, in terms of sample efficiency and generation capabilities. Finally, we discover that a separated critic plays an important role in exploration and data collection during training. Our code, trained models and data are accessible at https://github.com/francelico/deac-rep.

comment: Published as a conference paper at ICLR 2025. 10 pages

♻ ☆ ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery ICLR 2025

Ziru Chen, Shijie Chen, Yuting Ning, Qianheng Zhang, Boshi Wang, Botao Yu, Yifei Li, Zeyi Liao, Chen Wei, Zitong Lu, Vishal Dey, Mingyi Xue, Frazier N. Baker, Benjamin Burns, Daniel Adu-Ampratwum, Xuhui Huang, Xia Ning, Song Gao, Yu Su, Huan Sun

The advancements of large language models (LLMs) have piqued growing interest in developing LLM-based language agents to automate scientific discovery end-to-end, which has sparked both excitement and skepticism about their true capabilities. In this work, we call for rigorous assessment of agents on individual tasks in a scientific workflow before making bold claims on end-to-end automation. To this end, we present ScienceAgentBench, a new benchmark for evaluating language agents for data-driven scientific discovery. To ensure the scientific authenticity and real-world relevance of our benchmark, we extract 102 tasks from 44 peer-reviewed publications in four disciplines and engage nine subject matter experts to validate them. We unify the target output for every task to a self-contained Python program file and employ an array of evaluation metrics to examine the generated programs, execution results, and costs. Each task goes through multiple rounds of manual validation by annotators and subject matter experts to ensure its annotation quality and scientific plausibility. We also propose two effective strategies to mitigate data contamination concerns. Using ScienceAgentBench, we evaluate five open-weight and proprietary LLMs, each with three frameworks: direct prompting, OpenHands CodeAct, and self-debug. Given three attempts for each task, the best-performing agent can only solve 32.4% of the tasks independently and 34.3% with expert-provided knowledge. In addition, we evaluate OpenAI o1-preview with direct prompting and self-debug, which can boost the performance to 42.2%, demonstrating the effectiveness of increasing inference-time compute but with more than 10 times the cost of other LLMs. Still, our results underscore the limitations of current language agents in generating code for data-driven discovery, let alone end-to-end automation for scientific research.

comment: ICLR 2025. 60 pages

♻ ☆ Concept Navigation and Classification via Open-Source Large Language Model Processing

Maël Kubli

This paper presents a novel methodological framework for detecting and classifying latent constructs, including frames, narratives, and topics, from textual data using Open-Source Large Language Models (LLMs). The proposed hybrid approach combines automated summarization with human-in-the-loop validation to enhance the accuracy and interpretability of construct identification. By employing iterative sampling coupled with expert refinement, the framework guarantees methodological robustness and ensures conceptual precision. Applied to diverse data sets, including AI policy debates, newspaper articles on encryption, and the 20 Newsgroups data set, this approach demonstrates its versatility in systematically analyzing complex political discourses, media framing, and topic classification tasks.

comment: 36 pages, 1 figure, 5 tabels

♻ ☆ PhD Knowledge Not Required: A Reasoning Challenge for Large Language Models

Zixuan Wu, Francesca Lucchetti, Aleksander Boruch-Gruszecki, Jingmiao Zhao, Carolyn Jane Anderson, Joydeep Biswas, Federico Cassano, Molly Q Feldman, Arjun Guha

Existing benchmarks for frontier models often test specialized, "PhD-level" knowledge that is difficult for non-experts to grasp. In contrast, we present a benchmark with 594 problems based on the NPR Sunday Puzzle Challenge that requires only general knowledge. Our benchmark is challenging for both humans and models; however correct solutions are easy to verify, and models' mistakes are easy to spot. As LLMs are more widely deployed in society, we believe it is useful to develop benchmarks for frontier models that humans can understand without the need for deep domain expertise. Our work reveals capability gaps that are not evident in existing benchmarks: OpenAI o1 significantly outperforms other reasoning models on our benchmark, despite being on par with other models when tested on benchmarks that test specialized knowledge. Furthermore, our analysis of reasoning outputs uncovers new kinds of failures. DeepSeek R1, for instance, often concedes with "I give up" before providing an answer that it knows is wrong. R1 can also be remarkably "uncertain" in its output and in rare cases, it does not "finish thinking," which suggests the need for techniques to "wrap up" before the context window limit is reached. We also quantify the effectiveness of reasoning longer to identify the point beyond which more reasoning is unlikely to improve accuracy on our benchmark.

♻ ☆ Backdoor Graph Condensation ICDE 2025

Jiahao Wu, Ning Lu, Zeiyu Dai, Kun Wang, Wenqi Fan, Shengcai Liu, Qing Li, Ke Tang

Graph condensation has recently emerged as a prevalent technique to improve the training efficiency for graph neural networks (GNNs). It condenses a large graph into a small one such that a GNN trained on this small synthetic graph can achieve comparable performance to a GNN trained on the large graph. However, while existing graph condensation studies mainly focus on the best trade-off between graph size and the GNNs' performance (model utility), they overlook the security issues of graph condensation. To bridge this gap, we first explore backdoor attack against the GNNs trained on the condensed graphs. We introduce an effective backdoor attack against graph condensation, termed BGC. This attack aims to (1) preserve the condensed graph quality despite trigger injection, and (2) ensure trigger efficacy through the condensation process, achieving a high attack success rate. Specifically, BGC consistently updates triggers during condensation and targets representative nodes for poisoning. Extensive experiments demonstrate the effectiveness of our attack. BGC achieves a high attack success rate (close to 1.0) and good model utility in all cases. Furthermore, the results against multiple defense methods demonstrate BGC's resilience under their defenses. Finally, we analyze the key hyperparameters that influence the attack performance. Our code is available at: https://github.com/JiahaoWuGit/BGC.

comment: ICDE 2025 Camera Ready

♻ ☆ Innovative LSGTime Model for Crime Spatiotemporal Prediction Based on MindSpore Framework

Zhenkai Qin, BaoZhong Wei, Caifeng Gao

With the acceleration of urbanization, the spatiotemporal characteristics of criminal activities have become increasingly complex. Accurate prediction of crime distribution is crucial for optimizing the allocation of police resources and preventing crime. This paper proposes LGSTime, a crime spatiotemporal prediction model that integrates Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and the Multi-head Sparse Self-attention mechanism. LSTM and GRU capture long-term dependencies in crime time series, such as seasonality and periodicity, through their unique gating mechanisms. The Multi-head Sparse Self-attention mechanism, on the other hand, focuses on both temporal and spatial features of criminal events simultaneously through parallel processing and sparsification techniques, significantly improving computational efficiency and prediction accuracy. The integrated model leverages the strengths of each technique to better handle complex spatiotemporal data. Experimental findings demonstrate that the model attains optimal performance across four real - world crime datasets. In comparison to the CNN model, it exhibits performance enhancements of 2.8\%, 1.9\%, and 1.4\% in the Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) metrics respectively. These results offer a valuable reference for tackling the challenges in crime prediction.

♻ ☆ AI in radiological imaging of soft-tissue and bone tumours: a systematic review evaluating against CLAIM and FUTURE-AI guidelines

Douwe J. Spaanderman, Matthew Marzetti, Xinyi Wan, Andrew F. Scarsbrook, Philip Robinson, Edwin H. G. Oei, Jacob J. Visser, Robert Hemke, Kirsten van Langevelde, David F. Hanff, Geert J. L. H. van Leenders, Cornelis Verhoef, Dirk J. Gruühagen, Wiro J. Niessen, Stefan Klein, Martijn P. A. Starmans

Soft-tissue and bone tumours (STBT) are rare, diagnostically challenging lesions with variable clinical behaviours and treatment approaches. This systematic review provides an overview of Artificial Intelligence (AI) methods using radiological imaging for diagnosis and prognosis of these tumours, highlighting challenges in clinical translation, and evaluating study alignment with the Checklist for AI in Medical Imaging (CLAIM) and the FUTURE-AI international consensus guidelines for trustworthy and deployable AI to promote the clinical translation of AI methods. The review covered literature from several bibliographic databases, including papers published before 17/07/2024. Original research in peer-reviewed journals focused on radiology-based AI for diagnosing or prognosing primary STBT was included. Exclusion criteria were animal, cadaveric, or laboratory studies, and non-English papers. Abstracts were screened by two of three independent reviewers for eligibility. Eligible papers were assessed against guidelines by one of three independent reviewers. The search identified 15,015 abstracts, from which 325 articles were included for evaluation. Most studies performed moderately on CLAIM, averaging a score of 28.9$\pm$7.5 out of 53, but poorly on FUTURE-AI, averaging 5.1$\pm$2.1 out of 30. Imaging-AI tools for STBT remain at the proof-of-concept stage, indicating significant room for improvement. Future efforts by AI developers should focus on design (e.g. define unmet clinical need, intended clinical setting and how AI would be integrated in clinical workflow), development (e.g. build on previous work, explainability), evaluation (e.g. evaluating and addressing biases, evaluating AI against best practices), and data reproducibility and availability (making documented code and data publicly available). Following these recommendations could improve clinical translation of AI methods.

comment: 25 pages, 6 figures, 8 supplementary figures

♻ ☆ Score-Based Metropolis-Hastings Algorithms

Ahmed Aloui, Ali Hasan, Juncheng Dong, Zihao Wu, Vahid Tarokh

In this paper, we introduce a new approach for integrating score-based models with the Metropolis-Hastings algorithm. While traditional score-based diffusion models excel in accurately learning the score function from data points, they lack an energy function, making the Metropolis-Hastings adjustment step inaccessible. Consequently, the unadjusted Langevin algorithm is often used for sampling using estimated score functions. The lack of an energy function then prevents the application of the Metropolis-adjusted Langevin algorithm and other Metropolis-Hastings methods, limiting the wealth of other algorithms developed that use acceptance functions. We address this limitation by introducing a new loss function based on the \emph{detailed balance condition}, allowing the estimation of the Metropolis-Hastings acceptance probabilities given a learned score function. We demonstrate the effectiveness of the proposed method for various scenarios, including sampling from heavy-tail distributions.

♻ ☆ Learning Beamforming Codebooks for Active Sensing with Reconfigurable Intelligent Surface

Zhongze Zhang, Wei Yu

This paper explores the design of beamforming codebooks for the base station (BS) and for the reconfigurable intelligent surfaces (RISs) in an active sensing scheme for uplink localization, in which the mobile user transmits a sequence of pilots to the BS through reflection at the RISs, and the BS and the RISs are adaptively configured by carefully choosing BS beamforming codeword and RIS codewords from their respective codebooks in a sequential manner to progressively focus onto the user. Most existing codebook designs for RIS are not tailored for active sensing, by which we mean the choice of the next codeword should depend on the measurements made so far, and the sequence of codewords should dynamically focus reflection toward the user. Moreover, most existing codeword selection methods rely on exhaustive search in beam training to identify the codeword with the highest signal-to-noise ratio (SNR), thus incurring substantial pilot overhead as the size of the codebook scales. This paper proposes a learning-based approach for codebook construction and for codeword selection for active sensing. The proposed learning approach aims to locate a target in the service area by recursively selecting a sequence of BS beamforming codewords and RIS codewords from the respective codebooks as more measurements become available without exhaustive beam training. The codebook design and the codeword selection fuse key ideas from the vector quantized variational autoencoder (VQ-VAE) and the long short-term memory (LSTM) network to learn respectively the discrete function space of the codebook and the temporal dependencies between measurements.

comment: Accepted in IEEE Transactions on Wireless Communications

♻ ☆ Exploring Data Scaling Trends and Effects in Reinforcement Learning from Human Feedback

Wei Shen, Guanlin Liu, Zheng Wu, Ruofei Zhu, Qingping Yang, Chao Xin, Yu Yue, Lin Yan

Reinforcement Learning from Human Feedback (RLHF) is crucial for aligning large language models with human preferences. While recent research has focused on algorithmic improvements, the importance of prompt-data construction has been overlooked. This paper addresses this gap by exploring data-driven bottlenecks in RLHF performance scaling, particularly reward hacking and decreasing response diversity. We introduce a hybrid reward system combining reasoning task verifiers (RTV) and a generative reward model (GenRM) to mitigate reward hacking. We also propose a novel prompt-selection method, Pre-PPO, to maintain response diversity and enhance learning effectiveness. Additionally, we find that prioritizing mathematical and coding tasks early in RLHF training significantly improves performance. Experiments across two model sizes validate our methods' effectiveness and scalability. Results show that RTV is most resistant to reward hacking, followed by GenRM with ground truth, and then GenRM with SFT Best-of-N responses. Our strategies enable rapid capture of subtle task-specific distinctions, leading to substantial improvements in overall RLHF performance. This work highlights the importance of careful data construction and provides practical methods to overcome performance barriers in RLHF.

♻ ☆ Efficient Learning for Entropy-Regularized Markov Decision Processes via Multilevel Monte Carlo

Matthieu Meunier, Christoph Reisinger, Yufei Zhang

Designing efficient learning algorithms with complexity guarantees for Markov decision processes (MDPs) with large or continuous state and action spaces remains a fundamental challenge. We address this challenge for entropy-regularized MDPs with Polish state and action spaces, assuming access to a generative model of the environment. We propose a novel family of multilevel Monte Carlo (MLMC) algorithms that integrate fixed-point iteration with MLMC techniques and a generic stochastic approximation of the Bellman operator. We quantify the precise impact of the chosen approximate Bellman operator on the accuracy of the resulting MLMC estimator. Leveraging this error analysis, we show that using a biased plain MC estimate for the Bellman operator results in quasi-polynomial sample complexity, whereas an unbiased randomized multilevel approximation of the Bellman operator achieves polynomial sample complexity in expectation. Notably, these complexity bounds are independent of the dimensions or cardinalities of the state and action spaces, distinguishing our approach from existing algorithms whose complexities scale with the sizes of these spaces. We validate these theoretical performance guarantees through numerical experiments.

comment: 46 pages, 6 figures; fixed formatting of definitions and titles

♻ ☆ Are Large Language Models Memorizing Bug Benchmarks?

Daniel Ramos, Claudia Mamede, Kush Jain, Paulo Canelas, Catarina Gamboa, Claire Le Goues

Large Language Models (LLMs) have become integral to various software engineering tasks, including code generation, bug detection, and repair. To evaluate model performance in these domains, numerous bug benchmarks containing real-world bugs from software projects have been developed. However, a growing concern within the software engineering community is that these benchmarks may not reliably reflect true LLM performance due to the risk of data leakage. Despite this concern, limited research has been conducted to quantify the impact of potential leakage. In this paper, we systematically evaluate popular LLMs to assess their susceptibility to data leakage from widely used bug benchmarks. To identify potential leakage, we use multiple metrics, including a study of benchmark membership within commonly used training datasets, as well as analyses of negative log-likelihood and n-gram accuracy. Our findings show that certain models, in particular codegen-multi, exhibit significant evidence of memorization in widely used benchmarks like Defects4J, while newer models trained on larger datasets like LLaMa 3.1 exhibit limited signs of leakage. These results highlight the need for careful benchmark selection and the adoption of robust metrics to adequately assess models capabilities.

♻ ☆ LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning CVPR 2025

Xuan Liu, Xiaobin Chang

In continual learning (CL), catastrophic forgetting often arises due to feature drift. This challenge is particularly prominent in the exemplar-free continual learning (EFCL) setting, where samples from previous tasks cannot be retained, making it difficult to preserve prior knowledge. To address this issue, some EFCL methods aim to identify feature spaces that minimize the impact on previous tasks while accommodating new ones. However, they rely on static features or outdated statistics stored from old tasks, which prevents them from capturing the dynamic evolution of the feature space in CL, leading to performance degradation over time. In this paper, we introduce the Drift-Resistant Space (DRS), which effectively handles feature drifts without requiring explicit feature modeling or the storage of previous tasks. A novel parameter-efficient fine-tuning approach called Low-Rank Adaptation Subtraction (LoRA-) is proposed to develop the DRS. This method subtracts the LoRA weights of old tasks from the initial pre-trained weight before processing new task data to establish the DRS for model training. Therefore, LoRA- enhances stability, improves efficiency, and simplifies implementation. Furthermore, stabilizing feature drifts allows for better plasticity by learning with a triplet loss. Our method consistently achieves state-of-the-art results, especially for long task sequences, across multiple datasets.

comment: Accepted to CVPR 2025

♻ ☆ Accelerated Smoothing: A Scalable Approach to Randomized Smoothing

Devansh Bhardwaj, Kshitiz Kaushik, Sarthak Gupta

Randomized smoothing has emerged as a potent certifiable defense against adversarial attacks by employing smoothing noises from specific distributions to ensure the robustness of a smoothed classifier. However, the utilization of Monte Carlo sampling in this process introduces a compute-intensive element, which constrains the practicality of randomized smoothing on a larger scale. To address this limitation, we propose a novel approach that replaces Monte Carlo sampling with the training of a surrogate neural network. Through extensive experimentation in various settings, we demonstrate the efficacy of our approach in approximating the smoothed classifier with remarkable precision. Furthermore, we demonstrate that our approach significantly accelerates the robust radius certification process, providing nearly $600$X improvement in computation time, overcoming the computational bottlenecks associated with traditional randomized smoothing.

♻ ☆ The Mathematical Relationship Between Layer Normalization and Dynamic Activation Functions

Felix Stollenwerk

A recent paper proposes Dynamic Tanh (DyT) as a drop-in replacement for layer normalization (LN). Although the method is empirically well-motivated and appealing from a practical point of view, it lacks a theoretical foundation. In this work, we shed light on the mathematical relationship between layer normalization and dynamic activation functions. In particular, we derive DyT from LN and show that a well-defined approximation is needed to do so. By dropping said approximation, an alternative activation function is obtained, which we call Dynamic Inverse Square Root Unit (DyISRU). DyISRU is the exact counterpart of layer normalization, and we demonstrate numerically that it indeed resembles LN more accurately than DyT does.

comment: New title, renamed DyISRU, added missing parentheses in proof of theorem 3, minor language corrections

♻ ☆ LSEAttention is All You Need for Time Series Forecasting

Dizhen Liang

Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines. Previous research has identified the traditional attention mechanism as a key factor limiting their effectiveness in this domain. To bridge this gap, we introduce LATST, a novel approach designed to mitigate entropy collapse and training instability common challenges in Transformer-based time series forecasting. We rigorously evaluate LATST across multiple real-world multivariate time series datasets, demonstrating its ability to outperform existing state-of-the-art Transformer models. Notably, LATST manages to achieve competitive performance with fewer parameters than some linear models on certain datasets, highlighting its efficiency and effectiveness.

comment: 8 pages with referencing, 1 figure, 5 tables

♻ ☆ SoftCVI: Contrastive variational inference with self-generated soft labels ICLR

Daniel Ward, Mark Beaumont, Matteo Fasiolo

Estimating a distribution given access to its unnormalized density is pivotal in Bayesian inference, where the posterior is generally known only up to an unknown normalizing constant. Variational inference and Markov chain Monte Carlo methods are the predominant tools for this task; however, both are often challenging to apply reliably, particularly when the posterior has complex geometry. Here, we introduce Soft Contrastive Variational Inference (SoftCVI), which allows a family of variational objectives to be derived through a contrastive estimation framework. The approach parameterizes a classifier in terms of a variational distribution, reframing the inference task as a contrastive estimation problem aiming to identify a single true posterior sample among a set of samples. Despite this framing, we do not require positive or negative samples, but rather learn by sampling the variational distribution and computing ground truth soft classification labels from the unnormalized posterior itself. The objectives have zero variance gradient when the variational approximation is exact, without the need for specialized gradient estimators. We empirically investigate the performance on a variety of Bayesian inference tasks, using both simple (e.g. normal) and expressive (normalizing flow) variational distributions. We find that SoftCVI can be used to form objectives which are stable to train and mass-covering, frequently outperforming inference with other variational approaches.

comment: Updated to match version accepted at ICLR

♻ ☆ Pharmolix-FM: All-Atom Foundation Models for Molecular Modeling and Generation

Yizhen Luo, Jiashuo Wang, Siqi Fan, Zaiqing Nie

Structural biology relies on accurate three-dimensional biomolecular structures to advance our understanding of biological functions, disease mechanisms, and therapeutics. While recent advances in deep learning have enabled the development of all-atom foundation models for molecular modeling and generation, existing approaches face challenges in generalization due to the multi-modal nature of atomic data and the lack of comprehensive analysis of training and sampling strategies. To address these limitations, we propose PharMolixFM, a unified framework for constructing all-atom foundation models based on multi-modal generative techniques. Our framework includes three variants using state-of-the-art multi-modal generative models. By formulating molecular tasks as a generalized denoising process with task-specific priors, PharMolixFM achieves robust performance across various structural biology applications. Experimental results demonstrate that PharMolixFM-Diff achieves competitive prediction accuracy in protein-small-molecule docking (83.9% vs. 90.2% RMSD < 2{\AA}, given pocket) with significantly improved inference speed. Moreover, we explore the empirical inference scaling law by introducing more sampling repeats or steps. Our code and model are available at https://github.com/PharMolix/OpenBioMed.

♻ ☆ RelChaNet: Neural Network Feature Selection using Relative Change Scores

Felix Zimmer

There is an ongoing effort to develop feature selection algorithms to improve interpretability, reduce computational resources, and minimize overfitting in predictive models. Neural networks stand out as architectures on which to build feature selection methods, and recently, neuron pruning and regrowth have emerged from the sparse neural network literature as promising new tools. We introduce RelChaNet, a novel and lightweight supervised feature selection algorithm that uses neuron pruning and regrowth in the input layer of a dense neural network. For neuron pruning, a gradient sum metric measures the relative change induced in a network after a feature enters, while neurons are randomly regrown. We also propose an extension that adapts the size of the input layer at runtime. Extensive experiments on 13 different datasets show that our approach generally outperforms the current state-of-the-art methods, and in particular improves the average accuracy by 2% on the MNIST dataset. Our code is available at https://github.com/flxzimmer/relchanet.

♻ ☆ Emergent representations in networks trained with the Forward-Forward algorithm

Niccolò Tosato, Lorenzo Basile, Emanuele Ballarin, Giuseppe de Alteriis, Alberto Cazzaniga, Alessio Ansuini

The Backpropagation algorithm has often been criticised for its lack of biological realism. In an attempt to find a more biologically plausible alternative, the recently introduced Forward-Forward algorithm replaces the forward and backward passes of Backpropagation with two forward passes. In this work, we show that the internal representations obtained by the Forward-Forward algorithm can organise into category-specific ensembles exhibiting high sparsity -- composed of a low number of active units. This situation is reminiscent of what has been observed in cortical sensory areas, where neuronal ensembles are suggested to serve as the functional building blocks for perception and action. Interestingly, while this sparse pattern does not typically arise in models trained with standard Backpropagation, it can emerge in networks trained with Backpropagation on the same objective proposed for the Forward-Forward algorithm.

comment: Published in Transactions on Machine Learning Research (TMLR)

♻ ☆ Entropy-guided sequence weighting for efficient exploration in RL-based LLM fine-tuning

Abdullah Vanlioglu

We introduce Entropy-Guided Sequence Weighting (EGSW), a novel approach that enhances the exploration-exploitation tradeoff by dynamically assigning weights to generated outputs based on their advantage and entropy for Reinforcement Learning-based Large Language Model fine-tuning. EGSW integrates entropy regularization with advantage-based weighting to balance policy updates, enabling efficient exploration in high-dimensional state spaces. By employing temperature-scaled softmax weighting over sequences, EGSW prioritizing high-reward, high-uncertainty steps while maintaining training stability. Although originally developed to improve Group Relative Policy Optimization (GRPO) during large language model (LLM) fine-tuning, EGSW is generalizable to other reinforcement learning (RL) algorithms and can be implemented in both step-wise and trajectory-wise settings. Empirical evaluations demonstrate that EGSW enhances GRPO reasoning ability, yielding improvements in sample efficiency. Future work will explore the application of EGSW to advanced RL methodologies.

♻ ☆ Grasping a Handful: Sequential Multi-Object Dexterous Grasp Generation

Haofei Lu, Yifei Dong, Zehang Weng, Jens Lundell, Danica Kragic

We introduce the sequential multi-object robotic grasp sampling algorithm SeqGrasp that can robustly synthesize stable grasps on diverse objects using the robotic hand's partial Degrees of Freedom (DoF). We use SeqGrasp to construct the large-scale Allegro Hand sequential grasping dataset SeqDataset and use it for training the diffusion-based sequential grasp generator SeqDiffuser. We experimentally evaluate SeqGrasp and SeqDiffuser against the state-of-the-art non-sequential multi-object grasp generation method MultiGrasp in simulation and on a real robot. The experimental results demonstrate that SeqGrasp and SeqDiffuser reach an 8.71%-43.33% higher grasp success rate than MultiGrasp. Furthermore, SeqDiffuser is approximately 1000 times faster at generating grasps than SeqGrasp and MultiGrasp.

comment: 8 pages, 7 figures

♻ ☆ Dynamic High-Order Control Barrier Functions with Diffuser for Safety-Critical Trajectory Planning at Signal-Free Intersections

Di Chen, Ruiguo Zhong, Kehua Chen, Zhiwei Shang, Meixin Zhu, Edward Chung

Planning safe and efficient trajectories through signal-free intersections presents significant challenges for autonomous vehicles (AVs), particularly in dynamic, multi-task environments with unpredictable interactions and an increased possibility of conflicts. This study aims to address these challenges by developing a unified, robust, adaptive framework to ensure safety and efficiency across three distinct intersection movements: left-turn, right-turn, and straight-ahead. Existing methods often struggle to reliably ensure safety and effectively learn multi-task behaviors from demonstrations in such environments. This study proposes a safety-critical planning method that integrates Dynamic High-Order Control Barrier Functions (DHOCBF) with a diffusion-based model, called Dynamic Safety-Critical Diffuser (DSC-Diffuser). The DSC-Diffuser leverages task-guided planning to enhance efficiency, allowing the simultaneous learning of multiple driving tasks from real-world expert demonstrations. Moreover, the incorporation of goal-oriented constraints significantly reduces displacement errors, ensuring precise trajectory execution. To further ensure driving safety in dynamic environments, the proposed DHOCBF framework dynamically adjusts to account for the movements of surrounding vehicles, offering enhanced adaptability and reduce the conservatism compared to traditional control barrier functions. Validity evaluations of DHOCBF, conducted through numerical simulations, demonstrate its robustness in adapting to variations in obstacle velocities, sizes, uncertainties, and locations, effectively maintaining driving safety across a wide range of complex and uncertain scenarios. Comprehensive performance evaluations demonstrate that DSC-Diffuser generates realistic, stable, and generalizable policies, providing flexibility and reliable safety assurance in complex multi-task driving scenarios.

comment: 11 figures, 5 tables, 15 pages

♻ ☆ The AI off-switch problem as a signalling game: bounded rationality and incomparability

Alessio Benavoli, Alessandro Facchini, Marco Zaffalon

The off-switch problem is a critical challenge in AI control: if an AI system resists being switched off, it poses a significant risk. In this paper, we model the off-switch problem as a signalling game, where a human decision-maker communicates its preferences about some underlying decision problem to an AI agent, which then selects actions to maximise the human's utility. We assume that the human is a bounded rational agent and explore various bounded rationality mechanisms. Using real machine learning models, we reprove prior results and demonstrate that a necessary condition for an AI system to refrain from disabling its off-switch is its uncertainty about the human's utility. We also analyse how message costs influence optimal strategies and extend the analysis to scenarios involving incomparability.

♻ ☆ Mitigating Covariate Shift in Imitation Learning for Autonomous Vehicles Using Latent Space Generative World Models ICRA 2025

Alexander Popov, Alperen Degirmenci, David Wehr, Shashank Hegde, Ryan Oldja, Alexey Kamenev, Bertrand Douillard, David Nistér, Urs Muller, Ruchi Bhargava, Stan Birchfield, Nikolai Smolyanskiy

We propose the use of latent space generative world models to address the covariate shift problem in autonomous driving. A world model is a neural network capable of predicting an agent's next state given past states and actions. By leveraging a world model during training, the driving policy effectively mitigates covariate shift without requiring an excessive amount of training data. During end-to-end training, our policy learns how to recover from errors by aligning with states observed in human demonstrations, so that at runtime it can recover from perturbations outside the training distribution. Additionally, we introduce a novel transformer-based perception encoder that employs multi-view cross-attention and a learned scene query. We present qualitative and quantitative results, demonstrating significant improvements upon prior state of the art in closed-loop testing in the CARLA simulator, as well as showing the ability to handle perturbations in both CARLA and NVIDIA's DRIVE Sim.

comment: 8 pages, 6 figures, updated in March 2025, original published in September 2024, for ICRA 2025 submission, for associated video file, see https://youtu.be/7m3bXzlVQvU

♻ ☆ ShapG: new feature importance method based on the Shapley value

Chi Zhao, Jing Liu, Elena Parilina

With wide application of Artificial Intelligence (AI), it has become particularly important to make decisions of AI systems explainable and transparent. In this paper, we proposed a new Explainable Artificial Intelligence (XAI) method called ShapG (Explanations based on Shapley value for Graphs) for measuring feature importance. ShapG is a model-agnostic global explanation method. At the first stage, it defines an undirected graph based on the dataset, where nodes represent features and edges are added based on calculation of correlation coefficients between features. At the second stage, it calculates an approximated Shapley value by sampling the data taking into account this graph structure. The sampling approach of ShapG allows to calculate the importance of features efficiently, i.e. to reduce computational complexity. Comparison of ShapG with other existing XAI methods shows that it provides more accurate explanations for two examined datasets. We also compared other XAI methods developed based on cooperative game theory with ShapG in running time, and the results show that ShapG exhibits obvious advantages in its running time, which further proves efficiency of ShapG. In addition, extensive experiments demonstrate a wide range of applicability of the ShapG method for explaining complex models. We find ShapG an important tool in improving explainability and transparency of AI systems and believe it can be widely used in various fields.

comment: This paper has been published in the journal "Engineering Applications of Artificial Intelligence"

♻ ☆ Quantifying the Capability Boundary of DeepSeek Models: An Application-Driven Performance Analysis

Kaikai Zhao, Zhaoxiang Liu, Xuejiao Lei, Jiaojiao Zhao, Zhenhong Long, Zipeng Wang, Ning Wang, Meijuan An, Qingliang Meng, Peijun Yang, Minjie Hua, Chaoyang Ma, Wen Liu, Kai Wang, Shiguo Lian

DeepSeek-R1, known for its low training cost and exceptional reasoning capabilities, has achieved state-of-the-art performance on various benchmarks. However, detailed evaluations for DeepSeek Series models from the perspective of real-world applications are lacking, making it challenging for users to select the most suitable DeepSeek models for their specific needs. To address this gap, we conduct a systematic evaluation of the DeepSeek-V3, DeepSeek-R1, DeepSeek-R1-Distill-Qwen series, DeepSeek-R1-Distill-Llama series, their corresponding 4-bit quantized models, and the reasoning model QwQ-32B using the enhanced A-Eval benchmark, A-Eval-2.0. Through a comparative analysis of original instruction-tuned models and their distilled counterparts, we investigate how reasoning enhancements impact performance across diverse practical tasks. To assist users in model selection, we quantify the capability boundary of DeepSeek models through performance tier classifications. Based on the quantification results, we develop a model selection handbook that clearly illustrates the relation among models, their capabilities and practical applications. This handbook enables users to select the most cost-effective models without efforts, ensuring optimal performance and resource efficiency in real-world applications. It should be noted that, despite our efforts to establish a comprehensive, objective, and authoritative evaluation benchmark, the selection of test samples, characteristics of data distribution, and the setting of evaluation criteria may inevitably introduce certain biases into the evaluation results. We will continuously optimize the evaluation benchmarks and periodically update this paper to provide more comprehensive and accurate evaluation results. Please refer to the latest version of the paper for the most current results and conclusions.

♻ ☆ FreqX: Analyze the Attribution Methods in Another Domain

Zechen Liu, Feiyang Zhang, Wei Song, Xiang Li, Wei Wei

Personalized Federal learning(PFL) allows clients to cooperatively train a personalized model without disclosing their private dataset. However, PFL suffers from Non-IID, heterogeneous devices, lack of fairness, and unclear contribution which urgently need the interpretability of deep learning model to overcome these challenges. These challenges proposed new demands for interpretability. Low cost, privacy, and detailed information. There is no current interpretability method satisfying them. In this paper, we propose a novel interpretability method \emph{FreqX} by introducing Signal Processing and Information Theory. Our experiments show that the explanation results of FreqX contain both attribution information and concept information. FreqX runs at least 10 times faster than the baselines which contain concept information.

comment: 16pages, 9 figures

♻ ☆ Implicit Bias and Fast Convergence Rates for Self-attention

Bhavya Vasudeva, Puneesh Deora, Christos Thrampoulidis

We study the fundamental optimization principles of self-attention, the defining mechanism of transformers, by analyzing the implicit bias of gradient-based optimizers in training a self-attention layer with a linear decoder in binary classification. Building on prior studies in linear logistic regression, recent findings demonstrate that the key-query matrix $W_t$ from gradient-descent (GD) converges in direction towards $W_{mm}$, which maximizes the margin between optimal and non-optimal tokens across sequences. However, this convergence is local, dependent on initial conditions, only holds asymptotically as the number of iterations increases, and leaves questions about the potential benefits of adaptive step-size rules unaddressed. To bridge this gap, we first establish scenarios for which convergence is provably \emph{global}. We then analyze two adaptive step-size strategies: normalized GD and Polyak step-size, demonstrating \emph{finite-time} convergence rates for $W_t$ to $W_{mm}$, and quantifying the sparsification rate of the attention map. These findings not only show that these strategies can accelerate parameter convergence over standard GD in a non-convex setting but also deepen the understanding of the implicit bias in self-attention, linking it more closely to the phenomena observed in linear logistic regression despite its intricate non-convex nature.

comment: Accepted in TMLR, 43 pages, 10 figures

♻ ☆ Q-fid: Quantum Circuit Fidelity Improvement with LSTM Networks

Yikai Mao, Shaswot Shresthamali, Masaaki Kondo

The fidelity of quantum circuits (QC) is influenced by several factors, including hardware characteristics, calibration status, and the transpilation process, all of which impact their susceptibility to noise. However, existing methods struggle to estimate and compare the noise performance of different circuit layouts due to fluctuating error rates and the absence of a standardized fidelity metric. In this work, Q-fid is introduced, a Long Short-Term Memory (LSTM) based fidelity prediction system accompanied by a novel metric designed to quantify the fidelity of quantum circuits. Q-fid provides an intuitive way to predict the noise performance of Noisy Intermediate-Scale Quantum (NISQ) circuits. This approach frames fidelity prediction as a Time Series Forecasting problem to analyze the tokenized circuits, capturing the causal dependence of the gate sequences and their impact on overall fidelity. Additionally, the model is capable of dynamically adapting to changes in hardware characteristics, ensuring accurate fidelity predictions under varying conditions. Q-fid achieves a high prediction accuracy with an average RMSE of 0.0515, up to 24.7x more accurate than the Qiskit transpile tool mapomatic. By offering a reliable method for fidelity prediction, Q-fid empowers developers to optimize transpilation strategies, leading to more efficient and noise-resilient quantum circuit implementations.

♻ ☆ Internet of Things-Based Smart Precision Farming in Soilless Agriculture:Opportunities and Challenges for Global Food Security

Monica Dutta, Deepali Gupta, Sumegh Tharewal, Deepam Goyal, Jasminder Kaur Sandhu, Manjit Kaur, Ahmad Ali Alzubi, Jazem Mutared Alanazi

The rapid growth of the global population and the continuous decline in cultivable land pose significant threats to food security. This challenge worsens as climate change further reduces the availability of farmland. Soilless agriculture, such as hydroponics, aeroponics, and aquaponics, offers a sustainable solution by enabling efficient crop cultivation in controlled environments. The integration of the Internet of Things (IoT) with smart precision farming improves resource efficiency, automates environmental control, and ensures stable and high-yield crop production. IoT-enabled smart farming systems utilize real-time monitoring, data-driven decision-making, and automation to optimize water and nutrient usage while minimizing human intervention. This paper explores the opportunities and challenges of IoT-based soilless farming, highlighting its role in sustainable agriculture, urban farming, and global food security. These advanced farming methods ensure greater productivity, resource conservation, and year-round cultivation. However, they also face challenges such as high initial investment, technological dependency, and energy consumption. Through a comprehensive study, bibliometric analysis, and comparative analysis, this research highlights current trends and research gaps. It also outlines future directions for researchers, policymakers, and industry stakeholders to drive innovation and scalability in IoT-driven soilless agriculture. By emphasizing the benefits of vertical farming and Controlled Environment Agriculture (CEA)-enabled soilless techniques, this paper supports informed decision-making to address food security challenges and promote sustainable agricultural innovations.

♻ ☆ Learning out-of-time-ordered correlators with classical kernel methods

John Tanner, Jason Pye, Jingbo Wang

Out-of-Time Ordered Correlators (OTOCs) are widely used to investigate information scrambling in quantum systems. However, directly computing OTOCs with classical computers is an expensive procedure. This is due to the need to classically simulate the dynamics of quantum many-body systems, which entails computational costs that scale rapidly with system size. Similarly, exact simulation of the dynamics with a quantum computer (QC) will either only be possible for short times with noisy intermediate-scale quantum (NISQ) devices, or will require a fault-tolerant QC which is currently beyond technological capabilities. This motivates a search for alternative approaches to determine OTOCs and related quantities. In this study, we explore four parameterised sets of Hamiltonians describing local one-dimensional quantum systems of interest in condensed matter physics. For each set, we investigate whether classical kernel methods (KMs) can accurately learn the XZ-OTOC and a particular sum of OTOCs, as functions of the Hamiltonian parameters. We frame the problem as a regression task, generating small batches of labelled data with classical tensor network methods for quantum many-body systems with up to 40 qubits. Using this data, we train a variety of standard kernel machines and observe that the Laplacian and radial basis function (RBF) kernels perform best, achieving a coefficient of determination (\(R^2\)) on the testing sets of at least 0.7167, with averages between 0.8112 and 0.9822 for the various sets of Hamiltonians, together with small root mean squared error and mean absolute error. Hence, after training, the models can replace further uses of tensor networks for calculating an OTOC function of a system within the parameterised sets. Accordingly, the proposed method can assist with extensive evaluations of an OTOC function.

comment: 19+ 18 pages, 6 figures, 14 tables

♻ ☆ Boost Your Human Image Generation Model via Direct Preference Optimization CVPR

Sanghyeon Na, Yonggyu Kim, Hyunjoon Lee

Human image generation is a key focus in image synthesis due to its broad applications, but even slight inaccuracies in anatomy, pose, or details can compromise realism. To address these challenges, we explore Direct Preference Optimization (DPO), which trains models to generate preferred (winning) images while diverging from non-preferred (losing) ones. However, conventional DPO methods use generated images as winning images, limiting realism. To overcome this limitation, we propose an enhanced DPO approach that incorporates high-quality real images as winning images, encouraging outputs to resemble real images rather than generated ones. However, implementing this concept is not a trivial task. Therefore, our approach, HG-DPO (Human image Generation through DPO), employs a novel curriculum learning framework that gradually improves the output of the model toward greater realism, making training more feasible. Furthermore, HG-DPO effectively adapts to personalized text-to-image tasks, generating high-quality and identity-specific images, which highlights the practical value of our approach.

comment: CVPR`2025

♻ ☆ Scalable Multi-modal Model Predictive Control via Duality-based Interaction Predictions

Hansung Kim, Siddharth H. Nair, Francesco Borrelli

We propose a hierarchical architecture designed for scalable real-time Model Predictive Control (MPC) in complex, multi-modal traffic scenarios. This architecture comprises two key components: 1) RAID-Net, a novel attention-based Recurrent Neural Network that predicts relevant interactions along the MPC prediction horizon between the autonomous vehicle and the surrounding vehicles using Lagrangian duality, and 2) a reduced Stochastic MPC problem that eliminates irrelevant collision avoidance constraints, enhancing computational efficiency. Our approach is demonstrated in a simulated traffic intersection with interactive surrounding vehicles, showcasing a 12x speed-up in solving the motion planning problem. A video demonstrating the proposed architecture in multiple complex traffic scenarios can be found here: https://youtu.be/-pRiOnPb9_c. GitHub: https://github.com/MPC-Berkeley/hmpc_raidnet

comment: Accepted at IEEE Intelligent Vehicles Symposium 2024

♻ ☆ Comparison of Metadata Representation Models for Knowledge Graph Embeddings

Shusaku Egami, Kyoumoto Matsushita, Takanori Ugai, Ken Fukuda

Hyper-relational Knowledge Graphs (HRKGs) extend traditional KGs beyond binary relations, enabling the representation of contextual, provenance, and temporal information in domains, such as historical events, sensor data, video content, and narratives. HRKGs can be structured using several Metadata Representation Models (MRMs), including Reification (REF), Singleton Property (SGP), and RDF-star (RDR). However, the effects of different MRMs on KG Embedding (KGE) and Link Prediction (LP) models remain unclear. This study evaluates MRMs in the context of LP tasks, identifies the limitations of existing evaluation frameworks, and introduces a new task that ensures fair comparisons across MRMs. Furthermore, we propose a framework that effectively reflects the knowledge representations of the three MRMs in latent space. Experiments on two types of datasets reveal that REF performs well in simple HRKGs, whereas SGP is less effective. However, in complex HRKGs, the differences among MRMs in the LP tasks are minimal. Our findings contribute to an optimal knowledge representation strategy for HRKGs in LP tasks.

comment: 11 pages, 9 Figures

♻ ☆ Testing Support Size More Efficiently Than Learning Histograms

Renato Ferreira Pinto Jr., Nathaniel Harms

Consider two problems about an unknown probability distribution $p$: 1. How many samples from $p$ are required to test if $p$ is supported on $n$ elements or not? Specifically, given samples from $p$, determine whether it is supported on at most $n$ elements, or it is "$\epsilon$-far" (in total variation distance) from being supported on $n$ elements. 2. Given $m$ samples from $p$, what is the largest lower bound on its support size that we can produce? The best known upper bound for problem (1) uses a general algorithm for learning the histogram of the distribution $p$, which requires $\Theta(\tfrac{n}{\epsilon^2 \log n})$ samples. We show that testing can be done more efficiently than learning the histogram, using only $O(\tfrac{n}{\epsilon \log n} \log(1/\epsilon))$ samples, nearly matching the best known lower bound of $\Omega(\tfrac{n}{\epsilon \log n})$. This algorithm also provides a better solution to problem (2), producing larger lower bounds on support size than what follows from previous work. The proof relies on an analysis of Chebyshev polynomial approximations outside the range where they are designed to be good approximations, and the paper is intended as an accessible self-contained exposition of the Chebyshev polynomial method.

comment: 40 pages. Minor edits, added Open questions

♻ ☆ Learning dynamical systems with hit-and-run random feature maps

Pinak Mandal, Georg A. Gottwald

We show how random feature maps can be used to forecast dynamical systems with excellent forecasting skill. We consider the tanh activation function and judiciously choose the internal weights in a data-driven manner such that the resulting features explore the nonlinear, non-saturated regions of the activation function. We introduce skip connections and construct a deep variant of random feature maps by combining several units. To mitigate the curse of dimensionality, we introduce localization where we learn local maps, employing conditional independence. Our modified random feature maps provide excellent forecasting skill for both single trajectory forecasts as well as long-time estimates of statistical properties, for a range of chaotic dynamical systems with dimensions up to 512. In contrast to other methods such as reservoir computers which require extensive hyperparameter tuning, we effectively need to tune only a single hyperparameter, and are able to achieve state-of-the-art forecast skill with much smaller networks.

♻ ☆ ADMM Algorithms for Residual Network Training: Convergence Analysis and Parallel Implementation

Jintao Xu, Yifei Li, Wenxun Xing

We propose both serial and parallel proximal (linearized) alternating direction method of multipliers (ADMM) algorithms for training residual neural networks. In contrast to backpropagation-based approaches, our methods inherently mitigate the exploding gradient issue and are well-suited for parallel and distributed training through regional updates. Theoretically, we prove that the proposed algorithms converge at an R-linear (sublinear) rate for both the iteration points and the objective function values. These results hold without imposing stringent constraints on network width, depth, or training data size. Furthermore, we theoretically analyze our parallel/distributed ADMM algorithms, highlighting their reduced time complexity and lower per-node memory consumption. To facilitate practical deployment, we develop a control protocol for parallel ADMM implementation using Python's multiprocessing and interprocess communication. Experimental results validate the proposed ADMM algorithms, demonstrating rapid and stable convergence, improved performance, and high computational efficiency. Finally, we highlight the improved scalability and efficiency achieved by our parallel ADMM training strategy.

♻ ☆ Training-Free Exponential Context Extension via Cascading KV Cache

Jeffrey Willette, Heejun Lee, Youngwan Lee, Myeongjae Jeon, Sung Ju Hwang

The transformer's context window is vital for tasks such as few-shot learning and conditional generation as it preserves previous tokens for active memory. However, as the context lengths increase, the computational costs grow quadratically, hindering the deployment of large language models (LLMs) in real-world, long sequence scenarios. Although some recent key-value caching (KV Cache) methods offer linear inference complexity, they naively manage the stored context, prematurely evicting tokens and losing valuable information. Moreover, they lack an optimized prefill/prompt stage strategy, resulting in higher latency than even quadratic attention for realistic context sizes. In response, we introduce a novel mechanism that leverages cascading sub-cache buffers to selectively retain the most relevant tokens, enabling the model to maintain longer context histories without increasing the cache size. Our approach outperforms linear caching baselines across key benchmarks, including streaming perplexity, question answering, book summarization, and passkey retrieval, where it retains better retrieval accuracy at 1M tokens after four doublings of the cache size of 65K. Additionally, our method reduces prefill stage latency by a factor of 6.8 when compared to flash attention on 1M tokens. These innovations not only enhance the computational efficiency of LLMs but also pave the way for their effective deployment in resource-constrained environments, enabling large-scale, real-time applications with significantly reduced latency.

♻ ☆ XAMBA: Enabling Efficient State Space Models on Resource-Constrained Neural Processing Units

Arghadip Das, Arnab Raha, Shamik Kundu, Soumendu Kumar Ghosh, Deepak Mathaikutty, Vijay Raghunathan

State-Space Models (SSMs) have emerged as efficient alternatives to transformers for sequential data tasks, offering linear or near-linear scalability with sequence length, making them ideal for long-sequence applications in NLP, vision, and edge AI, including real-time transcription, translation, and contextual search. These applications require lightweight, high-performance models for deployment on resource-constrained devices like laptops and PCs. Designing specialized accelerators for every emerging neural network is costly and impractical; instead, optimizing models for existing NPUs in AI PCs provides a scalable solution. To this end, we propose XAMBA, the first framework to enable and optimize SSMs on commercial off-the-shelf (COTS) state-of-the-art (SOTA) NPUs. XAMBA follows a three-step methodology: (1) enabling SSMs on NPUs, (2) optimizing performance to meet KPI requirements, and (3) trading accuracy for additional performance gains. After enabling SSMs on NPUs, XAMBA mitigates key bottlenecks using CumBA and ReduBA, replacing sequential CumSum and ReduceSum operations with matrix-based computations, significantly improving execution speed and memory efficiency. Additionally, ActiBA enhances performance by approximating expensive activation functions (e.g., Swish, Softplus) using piecewise linear mappings, reducing latency with minimal accuracy loss. Evaluations on an Intel Core Ultra Series 2 AI PC show that XAMBA achieves up to 4.8X speed-up over the baseline. Our implementation is available at https://github.com/arghadippurdue/XAMBA.

♻ ☆ Model Selection for Inverse Reinforcement Learning via Structural Risk Minimization

Chendi Qu, Jianping He, Xiaoming Duan, Jiming Chen

Inverse reinforcement learning (IRL) usually assumes the reward function model is pre-specified as a weighted sum of features and estimates the weighting parameters only. However, how to select features and determine a proper reward model is nontrivial and experience-dependent. A simplistic model is less likely to contain the ideal reward function, while a model with high complexity leads to substantial computation cost and potential overfitting. This paper addresses this trade-off in the model selection for IRL problems by introducing the structural risk minimization (SRM) framework from statistical learning. SRM selects an optimal reward function class from a hypothesis set minimizing both estimation error and model complexity. To formulate an SRM scheme for IRL, we estimate the policy gradient from given demonstration as the empirical risk, and establish the upper bound of Rademacher complexity as the model penalty of hypothesis function classes. The SRM learning guarantee is further presented. In particular, we provide the explicit form for the linear weighted sum setting. Simulations demonstrate the performance and efficiency of our algorithm.

♻ ☆ MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models ICLR 2025

Peng Xia, Siwei Han, Shi Qiu, Yiyang Zhou, Zhaoyang Wang, Wenhao Zheng, Zhaorun Chen, Chenhang Cui, Mingyu Ding, Linjie Li, Lijuan Wang, Huaxiu Yao

Interleaved multimodal comprehension and generation, enabling models to produce and interpret both images and text in arbitrary sequences, have become a pivotal area in multimodal learning. Despite significant advancements, the evaluation of this capability remains insufficient. Existing benchmarks suffer from limitations in data scale, scope, and evaluation depth, while current evaluation metrics are often costly or biased, lacking in reliability for practical applications. To address these challenges, we introduce MMIE, a large-scale knowledge-intensive benchmark for evaluating interleaved multimodal comprehension and generation in Large Vision-Language Models (LVLMs). MMIE comprises 20K meticulously curated multimodal queries, spanning 3 categories, 12 fields, and 102 subfields, including mathematics, coding, physics, literature, health, and arts. It supports both interleaved inputs and outputs, offering a mix of multiple-choice and open-ended question formats to evaluate diverse competencies. Moreover, we propose a reliable automated evaluation metric, leveraging a scoring model fine-tuned with human-annotated data and systematic evaluation criteria, aimed at reducing bias and improving evaluation accuracy. Extensive experiments demonstrate the effectiveness of our benchmark and metrics in providing a comprehensive evaluation of interleaved LVLMs. Specifically, we evaluate eight LVLMs, revealing that even the best models show significant room for improvement, with most achieving only moderate results. We believe MMIE will drive further advancements in the development of interleaved LVLMs. We publicly release our benchmark and code in https://mmie-bench.github.io/.

comment: ICLR 2025 Oral

♻ ☆ Tackling Copyright Issues in AI Image Generation Through Originality Estimation and Genericization

Hiroaki Chiba-Okabe, Weijie J. Su

The rapid progress of generative AI technology has sparked significant copyright concerns, leading to numerous lawsuits filed against AI developers. Notably, generative AI's capacity for generating images of copyrighted characters has been well documented in the literature, and while various techniques for mitigating copyright issues have been studied, significant risks remain. Here, we propose a genericization method that modifies the outputs of a generative model to make them more generic and less likely to imitate distinctive features of copyrighted materials. To achieve this, we introduce a metric for quantifying the level of originality of data, estimated by drawing samples from a generative model, and applied in the genericization process. As a practical implementation, we introduce PREGen (Prompt Rewriting-Enhanced Genericization), which combines our genericization method with an existing mitigation technique. Compared to the existing method, PREGen reduces the likelihood of generating copyrighted characters by more than half when the names of copyrighted characters are used as the prompt. Additionally, while generative models can produce copyrighted characters even when their names are not directly mentioned in the prompt, PREGen almost entirely prevents the generation of such characters in these cases. Ultimately, this study advances computational approaches for quantifying and strengthening copyright protection, thereby providing practical methodologies to promote responsible generative AI development.

comment: 23 pages, 10 figures

♻ ☆ Interpretable Few-shot Learning with Online Attribute Selection

Mohammad Reza Zarei, Majid Komeili

Few-shot learning (FSL) presents a challenging learning problem in which only a few samples are available for each class. Decision interpretation is more important in few-shot classification due to a greater chance of error compared to traditional classification. However, the majority of the previous FSL methods are black-box models. In this paper, we propose an inherently interpretable model for FSL based on human-friendly attributes. Previously, human-friendly attributes have been utilized to train models with the potential for human interaction and interpretability. However, such approaches are not directly extendible to the few-shot classification scenario. Moreover, we propose an online attribute selection mechanism to effectively filter out irrelevant attributes in each episode. The attribute selection mechanism improves accuracy and helps with interpretability by reducing the number of attributes that participate in each episode. We further propose a mechanism that automatically detects the episodes where the pool of available human-friendly attributes is insufficient, and subsequently augments it by engaging some learned unknown attributes. We demonstrate that the proposed method achieves results on par with black-box few-shot learning models on four widely used datasets. We also empirically evaluate the level of decision alignment between different models and human understanding and show that our model outperforms the comparison methods based on this criterion.

♻ ☆ ADMM for Structured Fractional Minimization

Ganzhao Yuan

This paper considers a class of structured fractional minimization problems. The numerator consists of a differentiable function, a simple nonconvex nonsmooth function, a concave nonsmooth function, and a convex nonsmooth function composed with a linear operator. The denominator is a continuous function that is either weakly convex or has a weakly convex square root. These problems are prevalent in various important applications in machine learning and data science. Existing methods, primarily based on subgradient methods and smoothing proximal gradient methods, often suffer from slow convergence and numerical stability issues. In this paper, we introduce {\sf FADMM}, the first Alternating Direction Method of Multipliers tailored for this class of problems. {\sf FADMM} decouples the original problem into linearized proximal subproblems, featuring two variants: one using Dinkelbach's parametric method ({\sf FADMM-D}) and the other using the quadratic transform method ({\sf FADMM-Q}). By introducing a novel Lyapunov function, we establish that {\sf FADMM} converges to $\epsilon$-approximate critical points of the problem within an oracle complexity of $\mathcal{O}(1/\epsilon^{3})$. Extensive experiments on synthetic and real-world datasets, including sparse Fisher discriminant analysis, robust Sharpe ratio minimization, and robust sparse recovery, demonstrate the effectiveness of our approach. Keywords: Fractional Minimization, Nonconvex Optimization, Proximal Linearized ADMM, Nonsmooth Optimization, Convergence Analysis

♻ ☆ A Minimal Control Family of Dynamical Systems for Universal Approximation

Yifei Duan, Yongqiang Cai

The universal approximation property (UAP) holds a fundamental position in deep learning, as it provides a theoretical foundation for the expressive power of neural networks. It is widely recognized that a composition of linear and nonlinear functions, such as the rectified linear unit (ReLU) activation function, can approximate continuous functions on compact domains. In this paper, we extend this efficacy to a scenario containing dynamical systems with controls. We prove that the control family $\mathcal{F}_1$ containing all affine maps and the nonlinear ReLU map is sufficient for generating flow maps that can approximate orientation-preserving (OP) diffeomorphisms on any compact domain. Since $\mathcal{F}_1$ contains only one nonlinear function and the UAP does not hold if we remove the nonlinear function, we call $\mathcal{F}_1$ a minimal control family for the UAP. On this basis, several mild sufficient conditions, such as affine invariance, are established for the control family and discussed. Our results reveal an underlying connection between the approximation power of neural networks and control systems and could provide theoretical guidance for examining the approximation power of flow-based models.

comment: 12 pages

♻ ☆ Self-Vocabularizing Training for Neural Machine Translation NAACL

Pin-Jie Lin, Ernie Chang

Past vocabulary learning techniques identify relevant vocabulary before training, relying on statistical and entropy-based assumptions that largely neglect the role of model training. Empirically, we observe that trained translation models are induced to use a byte-pair encoding (BPE) vocabulary subset distinct from the original BPE vocabulary, leading to performance improvements when retrained with the induced vocabulary. In this paper, we analyze this discrepancy in neural machine translation by examining vocabulary and entropy shifts during self-training--where each iteration generates a labeled dataset by pairing source sentences with the model's predictions to define a new vocabulary. Building on these insights, we propose self-vocabularizing training, an iterative method that self-selects a smaller, more optimal vocabulary, yielding up to a 1.49 BLEU improvement. Moreover, we find that deeper model architectures lead to both an increase in unique token usage and a 6-8% reduction in vocabulary size.

comment: Accepted to NAACL SRW 2025

♻ ☆ An Iterative Bayesian Approach for System Identification based on Linear Gaussian Models

Alexandros E. Tzikas, Mykel J. Kochenderfer

We tackle the problem of system identification, where we select inputs, observe the corresponding outputs from the true system, and optimize the parameters of our model to best fit the data. We propose a flexible and computationally tractable methodology that is compatible with any system and parametric family of models. Our approach only requires input-output data from the system and first-order information from the model with respect to the parameters. Our algorithm consists of two modules. First, we formulate the problem of system identification from a Bayesian perspective and use a linear Gaussian model approximation to iteratively optimize the model's parameters. In each iteration, we propose to use the input-output data to tune the covariance of the linear Gaussian model. This statistically calibrates the approach. Secondly, we define a Gaussian-based uncertainty measure for the model parameters, which we can then minimize with respect to the next selected input. We test our method with linear and nonlinear dynamics.

comment: Submitted to the IEEE CDC

♻ ☆ Graph neural networks extrapolate out-of-distribution for shortest paths

Robert R. Nerem, Samantha Chen, Sanjoy Dasgupta, Yusu Wang

Neural networks (NNs), despite their success and wide adoption, still struggle to extrapolate out-of-distribution (OOD), i.e., to inputs that are not well-represented by their training dataset. Addressing the OOD generalization gap is crucial when models are deployed in environments significantly different from the training set, such as applying Graph Neural Networks (GNNs) trained on small graphs to large, real-world graphs. One promising approach for achieving robust OOD generalization is the framework of neural algorithmic alignment, which incorporates ideas from classical algorithms by designing neural architectures that resemble specific algorithmic paradigms (e.g. dynamic programming). The hope is that trained models of this form would have superior OOD capabilities, in much the same way that classical algorithms work for all instances. We rigorously analyze the role of algorithmic alignment in achieving OOD generalization, focusing on graph neural networks (GNNs) applied to the canonical shortest path problem. We prove that GNNs, trained to minimize a sparsity-regularized loss over a small set of shortest path instances, exactly implement the Bellman-Ford (BF) algorithm for shortest paths. In fact, if a GNN minimizes this loss within an error of $\epsilon$, it implements the BF algorithm with an error of $O(\epsilon)$. Consequently, despite limited training data, these GNNs are guaranteed to extrapolate to arbitrary shortest-path problems, including instances of any size. Our empirical results support our theory by showing that NNs trained by gradient descent are able to minimize this loss and extrapolate in practice.

♻ ☆ Balls-and-Bins Sampling for DP-SGD AISTATS 2025

Lynn Chua, Badih Ghazi, Charlie Harrison, Ethan Leeman, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Amer Sinha, Chiyuan Zhang

We introduce the Balls-and-Bins sampling for differentially private (DP) optimization methods such as DP-SGD. While it has been common practice to use some form of shuffling in DP-SGD implementations, privacy accounting algorithms have typically assumed that Poisson subsampling is used instead. Recent work by Chua et al. (ICML 2024), however, pointed out that shuffling based DP-SGD can have a much larger privacy cost in practical regimes of parameters. In this work we show that the Balls-and-Bins sampling achieves the "best-of-both" samplers, namely, the implementation of Balls-and-Bins sampling is similar to that of Shuffling and models trained using DP-SGD with Balls-and-Bins sampling achieve utility comparable to those trained using DP-SGD with Shuffling at the same noise multiplier, and yet, Balls-and-Bins sampling enjoys similar-or-better privacy amplification as compared to Poisson subsampling in practical regimes.

comment: Conference Proceedings version for AISTATS 2025

♻ ☆ On Creating a Causally Grounded Usable Rating Method for Assessing the Robustness of Foundation Models Supporting Time Series

Kausik Lakkaraju, Rachneet Kaur, Parisa Zehtabi, Sunandita Patra, Siva Likitha Valluru, Zhen Zeng, Biplav Srivastava, Marco Valtorta

Foundation Models (FMs) have improved time series forecasting in various sectors, such as finance, but their vulnerability to input disturbances can hinder their adoption by stakeholders, such as investors and analysts. To address this, we propose a causally grounded rating framework to study the robustness of Foundational Models for Time Series (FMTS) with respect to input perturbations. We evaluate our approach to the stock price prediction problem, a well-studied problem with easily accessible public data, evaluating six state-of-the-art (some multi-modal) FMTS across six prominent stocks spanning three industries. The ratings proposed by our framework effectively assess the robustness of FMTS and also offer actionable insights for model selection and deployment. Within the scope of our study, we find that (1) multi-modal FMTS exhibit better robustness and accuracy compared to their uni-modal versions and, (2) FMTS pre-trained on time series forecasting task exhibit better robustness and forecasting accuracy compared to general-purpose FMTS pre-trained across diverse settings. Further, to validate our framework's usability, we conduct a user study showcasing FMTS prediction errors along with our computed ratings. The study confirmed that our ratings reduced the difficulty for users in comparing the robustness of different systems.

♻ ☆ A Formal Framework for Understanding Length Generalization in Transformers ICLR 2025

Xinting Huang, Andy Yang, Satwik Bhattamishra, Yash Sarrof, Andreas Krebs, Hattie Zhou, Preetum Nakkiran, Michael Hahn

A major challenge for transformers is generalizing to sequences longer than those observed during training. While previous works have empirically shown that transformers can either succeed or fail at length generalization depending on the task, theoretical understanding of this phenomenon remains limited. In this work, we introduce a rigorous theoretical framework to analyze length generalization in causal transformers with learnable absolute positional encodings. In particular, we characterize those functions that are identifiable in the limit from sufficiently long inputs with absolute positional encodings under an idealized inference scheme using a norm-based regularizer. This enables us to prove the possibility of length generalization for a rich family of problems. We experimentally validate the theory as a predictor of success and failure of length generalization across a range of algorithmic and formal language tasks. Our theory not only explains a broad set of empirical observations but also opens the way to provably predicting length generalization capabilities in transformers.

comment: 85 pages, 9 figures, 11 tables. Accepted for publication at ICLR 2025

♻ ☆ Diversity-driven Data Selection for Language Model Tuning through Sparse Autoencoder

Xianjun Yang, Shaoliang Nie, Lijuan Liu, Suchin Gururangan, Ujjwal Karn, Rui Hou, Madian Khabsa, Yuning Mao

Instruction tuning data are often quantity-saturated due to the large volume of data collection and fast model iteration, leaving data selection important but underexplored. Existing quality-driven data selection methods, such as LIMA (NeurIPS 2023 \citep{zhou2024lima}) and AlpaGasus (ICLR 2024 \citep{chenalpagasus}) generally ignore the equal importance of data diversity and complexity. In this work, we aim to design a diversity-aware data selection strategy and creatively propose using sparse autoencoders (SAEs) to tackle the challenge of data diversity measure. In addition, SAEs can also provide more interpretability of model behavior and explain, e.g., the surprising effectiveness of selecting the longest response (ICML 2024 \citep{zhaolong}). Using effective data selection, we experimentally prove that models trained on our selected data can outperform other methods in terms of model capabilities, reduce training cost, and potentially gain more control over model behaviors. We prove that SAEs can serve as a good alternative to diversity measure and design our method to be scalable for potential industrial large-scale pruning, and we will also release our trained SAEs for use by the broader community.

comment: fix typos

♻ ☆ Don't lie to your friends: Learning what you know from collaborative self-play

Jacob Eisenstein, Reza Aghajani, Adam Fisch, Dheeru Dua, Fantine Huot, Mirella Lapata, Vicky Zayats, Jonathan Berant

To be helpful assistants, AI agents must be aware of their own capabilities and limitations. This includes knowing when to answer from parametric knowledge versus using tools, when to trust tool outputs, and when to abstain or hedge. Such capabilities are hard to teach through supervised fine-tuning because they require constructing examples that reflect the agent's specific capabilities. We therefore propose a radically new approach to teaching agents what they know: \emph{collaborative self-play}. We construct multi-agent collaborations in which the group is rewarded for collectively arriving at correct answers. The desired meta-knowledge emerges from the incentives built into the structure of the interaction. We focus on small societies of agents that have access to heterogeneous tools (corpus-specific retrieval), and therefore must collaborate to maximize their success while minimizing their effort. Experiments show that group-level rewards for multi-agent communities can induce policies that \emph{transfer} to improve tool use and selective prediction in settings where individual agents are deployed in isolation.

♻ ☆ Towards Adversarially Robust Dataset Distillation by Curvature Regularization

Eric Xue, Yijiang Li, Haoyang Liu, Peiran Wang, Yifan Shen, Haohan Wang

Dataset distillation (DD) allows datasets to be distilled to fractions of their original size while preserving the rich distributional information so that models trained on the distilled datasets can achieve a comparable accuracy while saving significant computational loads. Recent research in this area has been focusing on improving the accuracy of models trained on distilled datasets. In this paper, we aim to explore a new perspective of DD. We study how to embed adversarial robustness in distilled datasets, so that models trained on these datasets maintain the high accuracy and meanwhile acquire better adversarial robustness. We propose a new method that achieves this goal by incorporating curvature regularization into the distillation process with much less computational overhead than standard adversarial training. Extensive empirical experiments suggest that our method not only outperforms standard adversarial training on both accuracy and robustness with less computation overhead but is also capable of generating robust distilled datasets that can withstand various adversarial attacks.

comment: 14 pages, 3 figures

♻ ☆ DT-DDNN: A Physical Layer Security Attack Detector in 5G RF Domain for CAVs

Ghazal Asemian, Mohammadreza Amini, Burak Kantarci, Melike Erol-Kantarci

The Synchronization Signal Block (SSB) is a fundamental component of the 5G New Radio (NR) air interface, crucial for the initial access procedure of Connected and Automated Vehicles (CAVs), and serves several key purposes in the network's operation. However, due to the predictable nature of SSB transmission, including the Primary and Secondary Synchronization Signals (PSS and SSS), jamming attacks are critical threats. These attacks, which can be executed without requiring high power or complex equipment, pose substantial risks to the 5G network, particularly as a result of the unencrypted transmission of control signals. Leveraging RF domain knowledge, this work presents a novel deep learning-based technique for detecting jammers in CAV networks. Unlike the existing jamming detection algorithms that mostly rely on network parameters, we introduce a double-threshold deep learning jamming detector by focusing on the SSB. The detection method is focused on RF domain features and improves the robustness of the network without requiring integration with the pre-existing network infrastructure. By integrating a preprocessing block to extract PSS correlation and energy per null resource elements (EPNRE) characteristics, our method distinguishes between normal and jammed received signals with high precision. Additionally, by incorporating of Discrete Wavelet Transform (DWT), the efficacy of training and detection are optimized. A double-threshold double Deep Neural Network (DT-DDNN) is also introduced to the architecture complemented by a deep cascade learning model to increase the sensitivity of the model to variations of signal-to-jamming noise ratio (SJNR). Results show that the proposed method achieves 96.4% detection rate in extra low jamming power, i.e., SJNR between 15 to 30 dB. Further, performance of DT-DDNN is validated by analyzing real 5G signals obtained from a practical testbed.

comment: 14 pages, 13 figures, accepted to IEEE Transactions on Vehicular Technology

♻ ☆ Learning Color Equivariant Representations ICLR 2025

Yulong Yang, Felix O'Mahony, Christine Allen-Blanchette

In this paper, we introduce group convolutional neural networks (GCNNs) equivariant to color variation. GCNNs have been designed for a variety of geometric transformations from 2D and 3D rotation groups, to semi-groups such as scale. Despite the improved interpretability, accuracy and generalizability of these architectures, GCNNs have seen limited application in the context of perceptual quantities. Notably, the recent CEConv network uses a GCNN to achieve equivariance to hue transformations by convolving input images with a hue rotated RGB filter. However, this approach leads to invalid RGB values which break equivariance and degrade performance. We resolve these issues with a lifting layer that transforms the input image directly, thereby circumventing the issue of invalid RGB values and improving equivariance error by over three orders of magnitude. Moreover, we extend the notion of color equivariance to include equivariance to saturation and luminance shift. Our hue-, saturation-, luminance- and color-equivariant networks achieve strong generalization to out-of-distribution perceptual variations and improved sample efficiency over conventional architectures. We demonstrate the utility of our approach on synthetic and real world datasets where we consistently outperform competitive baselines.

comment: Accept to The 13th International Conference on Learning Representations (ICLR 2025)

♻ ☆ Safe Policy Learning through Extrapolation: Application to Pre-trial Risk Assessment

Eli Ben-Michael, D. James Greiner, Kosuke Imai, Zhichao Jiang

Algorithmic recommendations and decisions have become ubiquitous in today's society. Many of these data-driven policies, especially in the realm of public policy, are based on known, deterministic rules to ensure their transparency and interpretability. We examine a particular case of algorithmic pre-trial risk assessments in the US criminal justice system, which provide deterministic classification scores and recommendations to help judges make release decisions. Our goal is to analyze data from a unique field experiment on an algorithmic pre-trial risk assessment to investigate whether the scores and recommendations can be improved. Unfortunately, prior methods for policy learning are not applicable because they require existing policies to be stochastic. We develop a maximin robust optimization approach that partially identifies the expected utility of a policy, and then finds a policy that maximizes the worst-case expected utility. The resulting policy has a statistical safety property, limiting the probability of producing a worse policy than the existing one, under structural assumptions about the outcomes. Our analysis of data from the field experiment shows that we can safely improve certain components of the risk assessment instrument by classifying arrestees as lower risk under a wide range of utility specifications, though the analysis is not informative about several components of the instrument.

♻ ☆ Eliminating Position Bias of Language Models: A Mechanistic Approach

Ziqi Wang, Hanlin Zhang, Xiner Li, Kuan-Hao Huang, Chi Han, Shuiwang Ji, Sham M. Kakade, Hao Peng, Heng Ji

Position bias has proven to be a prevalent issue of modern language models (LMs), where the models prioritize content based on its position within the given context. This bias often leads to unexpected model failures and hurts performance, robustness, and reliability across various applications. Our mechanistic analysis attributes the position bias to two components employed in nearly all state-of-the-art LMs: causal attention and relative positional encodings. Based on the analyses, we propose to eliminate position bias (e.g., different retrieved documents' orders in QA affect performance) with a training-free zero-shot approach. Our method changes the causal attention to bidirectional attention between documents and utilizes model attention values to decide the relative orders of documents instead of using the order provided in input prompts, therefore enabling Position-INvariant inferencE (PINE) at the document level. By eliminating position bias, models achieve better performance and reliability in downstream tasks, including LM-as-a-judge, retrieval-augmented QA, molecule generation, and math reasoning. Notably, PINE is especially useful when adapting LMs for evaluating reasoning pairs: it consistently provides 8 to 10 percentage points performance gains, making Llama-3-70B-Instruct perform even better than GPT-4-0125-preview and GPT-4o-2024-08-06 on the RewardBench reasoning set.

comment: 26 pages, 6 figures, 15 tables

♻ ☆ Features that Make a Difference: Leveraging Gradients for Improved Dictionary Learning NAACL 2025

Jeffrey Olmo, Jared Wilson, Max Forsey, Bryce Hepner, Thomas Vin Howe, David Wingate

Sparse Autoencoders (SAEs) are a promising approach for extracting neural network representations by learning a sparse and overcomplete decomposition of the network's internal activations. However, SAEs are traditionally trained considering only activation values and not the effect those activations have on downstream computations. This limits the information available to learn features, and biases the autoencoder towards neglecting features which are represented with small activation values but strongly influence model outputs. To address this, we introduce Gradient SAEs (g-SAEs), which modify the $k$-sparse autoencoder architecture by augmenting the TopK activation function to rely on the gradients of the input activation when selecting the $k$ elements. For a given sparsity level, g-SAEs produce reconstructions that are more faithful to original network performance when propagated through the network. Additionally, we find evidence that g-SAEs learn latents that are on average more effective at steering models in arbitrary contexts. By considering the downstream effects of activations, our approach leverages the dual nature of neural network features as both $\textit{representations}$, retrospectively, and $\textit{actions}$, prospectively. While previous methods have approached the problem of feature discovery primarily focused on the former aspect, g-SAEs represent a step towards accounting for the latter as well.

comment: 10 pages, 10 figures. Accepted to NAACL 2025

♻ ☆ VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models CVPR 2025

Muchao Ye, Weiyang Liu, Pan He

The rapid advancement of vision-language models (VLMs) has established a new paradigm in video anomaly detection (VAD): leveraging VLMs to simultaneously detect anomalies and provide comprehendible explanations for the decisions. Existing work in this direction often assumes the complex reasoning required for VAD exceeds the capabilities of pretrained VLMs. Consequently, these approaches either incorporate specialized reasoning modules during inference or rely on instruction tuning datasets through additional training to adapt VLMs for VAD. However, such strategies often incur substantial computational costs or data annotation overhead. To address these challenges in explainable VAD, we introduce a verbalized learning framework named VERA that enables VLMs to perform VAD without model parameter modifications. Specifically, VERA automatically decomposes the complex reasoning required for VAD into reflections on simpler, more focused guiding questions capturing distinct abnormal patterns. It treats these reflective questions as learnable parameters and optimizes them through data-driven verbal interactions between learner and optimizer VLMs, using coarsely labeled training data. During inference, VERA embeds the learned questions into model prompts to guide VLMs in generating segment-level anomaly scores, which are then refined into frame-level scores via the fusion of scene and temporal contexts. Experimental results on challenging benchmarks demonstrate that the learned questions of VERA are highly adaptable, significantly improving both detection performance and explainability of VLMs for VAD.

comment: Accepted in CVPR 2025

♻ ☆ VL-ICL Bench: The Devil in the Details of Multimodal In-Context Learning ICLR 2025

Yongshuo Zong, Ondrej Bohdal, Timothy Hospedales

Large language models (LLMs) famously exhibit emergent in-context learning (ICL) -- the ability to rapidly adapt to new tasks using few-shot examples provided as a prompt, without updating the model's weights. Built on top of LLMs, vision large language models (VLLMs) have advanced significantly in areas such as recognition, reasoning, and grounding. However, investigations into \emph{multimodal ICL} have predominantly focused on few-shot visual question answering (VQA), and image captioning, which we will show neither exploit the strengths of ICL, nor test its limitations. The broader capabilities and limitations of multimodal ICL remain under-explored. In this study, we introduce a comprehensive benchmark VL-ICL Bench for multimodal in-context learning, encompassing a broad spectrum of tasks that involve both images and text as inputs and outputs, and different types of challenges, from {perception to reasoning and long context length}. We evaluate the abilities of state-of-the-art VLLMs against this benchmark suite, revealing their diverse strengths and weaknesses, and showing that even the most advanced models, such as GPT-4, find the tasks challenging. By highlighting a range of new ICL tasks, and the associated strengths and limitations of existing models, we hope that our dataset will inspire future work on enhancing the in-context learning capabilities of VLLMs, as well as inspire new applications that leverage VLLM ICL. The code and dataset are available at https://github.com/ys-zong/VL-ICL.

comment: ICLR 2025

♻ ☆ Can Zero-Shot Commercial APIs Deliver Regulatory-Grade Clinical Text DeIdentification? ECIR 2025

Veysel Kocaman, Muhammed Santas, Yigit Gul, Mehmet Butgul, David Talby

We evaluate the performance of four leading solutions for de-identification of unstructured medical text - Azure Health Data Services, AWS Comprehend Medical, OpenAI GPT-4o, and John Snow Labs - on a ground truth dataset of 48 clinical documents annotated by medical experts. The analysis, conducted at both entity-level and token-level, suggests that John Snow Labs' Medical Language Models solution achieves the highest accuracy, with a 96% F1-score in protected health information (PHI) detection, outperforming Azure (91%), AWS (83%), and GPT-4o (79%). John Snow Labs is not only the only solution which achieves regulatory-grade accuracy (surpassing that of human experts) but is also the most cost-effective solution: It is over 80% cheaper compared to Azure and GPT-4o, and is the only solution not priced by token. Its fixed-cost local deployment model avoids the escalating per-request fees of cloud-based services, making it a scalable and economical choice.

comment: 14 pages, accepted at Text2Story Workshop at ECIR 2025

♻ ☆ Forgetting Transformer: Softmax Attention with a Forget Gate ICLR 2025

Zhixuan Lin, Evgenii Nikishin, Xu Owen He, Aaron Courville

An essential component of modern recurrent sequence models is the forget gate. While Transformers do not have an explicit recurrent form, we show that a forget gate can be naturally incorporated into Transformers by down-weighting the unnormalized attention scores in a data-dependent way. We name this attention mechanism Forgetting Attention and the resulting model the Forgetting Transformer (FoX). We show that FoX outperforms the Transformer on long-context language modeling, length extrapolation, and short-context downstream tasks, while performing on par with the Transformer on long-context downstream tasks. Moreover, it is compatible with the FlashAttention algorithm and does not require any positional embeddings. Several analyses, including the needle-in-the-haystack test, show that FoX also retains the Transformer's superior long-context capabilities over recurrent sequence models such as Mamba-2, HGRN2, and DeltaNet. We also introduce a "Pro" block design that incorporates some common architectural components in recurrent sequence models and find it significantly improves the performance of both FoX and the Transformer. Our code is available at https://github.com/zhixuan-lin/forgetting-transformer.

comment: Published as a conference paper at ICLR 2025; Fixed an issue with the attention map visualization

♻ ☆ SE Arena: An Interactive Platform for Evaluating Foundation Models in Software Engineering

Zhimin Zhao

Foundation models (FMs), particularly large language models (LLMs), have shown significant promise in various software engineering (SE) tasks, including code generation, debugging, and requirement refinement. Despite these advances, existing evaluation frameworks are insufficient for assessing model performance in iterative, context-rich workflows characteristic of SE activities. To address this limitation, we introduce SE Arena, an interactive platform designed to evaluate SE-focused chatbots. SE Arena provides a transparent, open-source leaderboard, supports multi-round conversational workflows, and enables end-to-end model comparisons. Moreover, SE Arena incorporates a new feature called RepoChat, which automatically injects repository-related context (e.g., issues, commits, pull requests) into the conversation, further aligning evaluations with real-world development processes. This paper outlines the design and capabilities of SE Arena, emphasizing its potential to advance the evaluation and practical application of FMs in software engineering.

comment: Check the arena at https://huggingface.co/spaces/SE-Arena/Software-Engineering-Arena

♻ ☆ Privacy Vulnerabilities in Marginals-based Synthetic Data

Steven Golob, Sikha Pentyala, Anuar Maratkhan, Martine De Cock

When acting as a privacy-enhancing technology, synthetic data generation (SDG) aims to maintain a resemblance to the real data while excluding personally-identifiable information. Many SDG algorithms provide robust differential privacy (DP) guarantees to this end. However, we show that the strongest class of SDG algorithms--those that preserve \textit{marginal probabilities}, or similar statistics, from the underlying data--leak information about individuals that can be recovered more efficiently than previously understood. We demonstrate this by presenting a novel membership inference attack, MAMA-MIA, and evaluate it against three seminal DP SDG algorithms: MST, PrivBayes, and Private-GSD. MAMA-MIA leverages knowledge of which SDG algorithm was used, allowing it to learn information about the hidden data more accurately, and orders-of-magnitude faster, than other leading attacks. We use MAMA-MIA to lend insight into existing SDG vulnerabilities. Our approach went on to win the first SNAKE (SaNitization Algorithm under attacK ... $\varepsilon$) competition.

comment: Accepted at 3rd IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) 2025

♻ ☆ Dataset Distillation via Knowledge Distillation: Towards Efficient Self-Supervised Pre-Training of Deep Networks ICLR 2025

Siddharth Joshi, Jiayi Ni, Baharan Mirzasoleiman

Dataset distillation (DD) generates small synthetic datasets that can efficiently train deep networks with a limited amount of memory and compute. Despite the success of DD methods for supervised learning, DD for self-supervised pre-training of deep models has remained unaddressed. Pre-training on unlabeled data is crucial for efficiently generalizing to downstream tasks with limited labeled data. In this work, we propose the first effective DD method for SSL pre-training. First, we show, theoretically and empirically, that naive application of supervised DD methods to SSL fails, due to the high variance of the SSL gradient. Then, we address this issue by relying on insights from knowledge distillation (KD) literature. Specifically, we train a small student model to match the representations of a larger teacher model trained with SSL. Then, we generate a small synthetic dataset by matching the training trajectories of the student models. As the KD objective has considerably lower variance than SSL, our approach can generate synthetic datasets that can successfully pre-train high-quality encoders. Through extensive experiments, we show that our distilled sets lead to up to 13% higher accuracy than prior work, on a variety of downstream tasks, in the presence of limited labeled data. Code at https://github.com/BigML-CS-UCLA/MKDT.

comment: ICLR 2025. Code at https://github.com/BigML-CS-UCLA/MKDT

♻ ☆ GP-MoLFormer: A Foundation Model For Molecular Generation

Jerret Ross, Brian Belgodere, Samuel C. Hoffman, Vijil Chenthamarakshan, Jiri Navratil, Youssef Mroueh, Payel Das

Transformer-based models trained on large and general purpose datasets consisting of molecular strings have recently emerged as a powerful tool for successfully modeling various structure-property relations. Inspired by this success, we extend the paradigm of training chemical language transformers on large-scale chemical datasets to generative tasks in this work. Specifically, we propose GP-MoLFormer, an autoregressive molecular string generator that is trained on more than 1.1B (billion) chemical SMILES. GP-MoLFormer uses a 46.8M parameter transformer decoder model with linear attention and rotary positional encodings as the base architecture. GP-MoLFormer's utility is evaluated and compared with that of existing baselines on three different tasks: de novo generation, scaffold-constrained molecular decoration, and unconstrained property-guided optimization. While the first two are handled with no additional training, we propose a parameter-efficient fine-tuning method for the last task, which uses property-ordered molecular pairs as input. We call this new approach pair-tuning. Our results show GP-MoLFormer performs better or comparable with baselines across all three tasks, demonstrating its general utility for a variety of molecular generation tasks. We further report strong memorization of training data in GP-MoLFormer generations, which has so far remained unexplored for chemical language models. Our analyses reveal that training data memorization and novelty in generations are impacted by the quality and scale of the training data; duplication bias in training data can enhance memorization at the cost of lowering novelty. We further establish a scaling law relating inference compute and novelty in generations.

♻ ☆ Severing Spurious Correlations with Data Pruning ICLR 2025

Varun Mulchandani, Jung-Eun Kim

Deep neural networks have been shown to learn and rely on spurious correlations present in the data that they are trained on. Reliance on such correlations can cause these networks to malfunction when deployed in the real world, where these correlations may no longer hold. To overcome the learning of and reliance on such correlations, recent studies propose approaches that yield promising results. These works, however, study settings where the strength of the spurious signal is significantly greater than that of the core, invariant signal, making it easier to detect the presence of spurious features in individual training samples and allow for further processing. In this paper, we identify new settings where the strength of the spurious signal is relatively weaker, making it difficult to detect any spurious information while continuing to have catastrophic consequences. We also discover that spurious correlations are learned primarily due to only a handful of all the samples containing the spurious feature and develop a novel data pruning technique that identifies and prunes small subsets of the training data that contain these samples. Our proposed technique does not require inferred domain knowledge, information regarding the sample-wise presence or nature of spurious information, or human intervention. Finally, we show that such data pruning attains state-of-the-art performance on previously studied settings where spurious information is identifiable.

comment: ICLR 2025, Spotlight

♻ ☆ Adversarially Robust Learning with Optimal Transport Regularized Divergences

Jeremiah Birrell, Reza Ebrahimi

We introduce a new class of optimal-transport-regularized divergences, $D^c$, constructed via an infimal convolution between an information divergence, $D$, and an optimal-transport (OT) cost, $C$, and study their use in distributionally robust optimization (DRO). In particular, we propose the $ARMOR_D$ methods as novel approaches to enhancing the adversarial robustness of deep learning models. These DRO-based methods are defined by minimizing the maximum expected loss over a $D^c$-neighborhood of the empirical distribution of the training data. Viewed as a tool for constructing adversarial samples, our method allows samples to be both transported, according to the OT cost, and re-weighted, according to the information divergence; the addition of a principled and dynamical adversarial re-weighting on top of adversarial sample transport is a key innovation of $ARMOR_D$. $ARMOR_D$ can be viewed as a generalization of the best-performing loss functions and OT costs in the adversarial training literature; we demonstrate this flexibility by using $ARMOR_D$ to augment the UDR, TRADES, and MART methods and obtain improved performance on CIFAR-10 and CIFAR-100 image recognition. Specifically, augmenting with $ARMOR_D$ leads to 1.9\% and 2.1\% improvement against AutoAttack, a powerful ensemble of adversarial attacks, on CIFAR-10 and CIFAR-100 respectively. To foster reproducibility, we made the code accessible at https://github.com/star-ailab/ARMOR.

comment: 33 pages, 2 figures

♻ ☆ Decentralized Bilevel Optimization: A Perspective from Transient Iteration Complexity

Boao Kong, Shuchen Zhu, Songtao Lu, Xinmeng Huang, Kun Yuan

Stochastic bilevel optimization (SBO) is becoming increasingly essential in machine learning due to its versatility in handling nested structures. To address large-scale SBO, decentralized approaches have emerged as effective paradigms in which nodes communicate with immediate neighbors without a central server, thereby improving communication efficiency and enhancing algorithmic robustness. However, most decentralized SBO algorithms focus solely on asymptotic convergence rates, overlooking transient iteration complexity-the number of iterations required before asymptotic rates dominate, which results in limited understanding of the influence of network topology, data heterogeneity, and the nested bilevel algorithmic structures. To address this issue, this paper introduces D-SOBA, a Decentralized Stochastic One-loop Bilevel Algorithm framework. D-SOBA comprises two variants: D-SOBA-SO, which incorporates second-order Hessian and Jacobian matrices, and D-SOBA-FO, which relies entirely on first-order gradients. We provide a comprehensive non-asymptotic convergence analysis and establish the transient iteration complexity of D-SOBA. This provides the first theoretical understanding of how network topology, data heterogeneity, and nested bilevel structures influence decentralized SBO. Extensive experimental results demonstrate the efficiency and theoretical advantages of D-SOBA.

comment: 59 pages, 7 figures

♻ ☆ Online Reinforcement Learning in Non-Stationary Context-Driven Environments ICLR '25

Pouya Hamadanian, Arash Nasr-Esfahany, Malte Schwarzkopf, Siddartha Sen, Mohammad Alizadeh

We study online reinforcement learning (RL) in non-stationary environments, where a time-varying exogenous context process affects the environment dynamics. Online RL is challenging in such environments due to "catastrophic forgetting" (CF). The agent tends to forget prior knowledge as it trains on new experiences. Prior approaches to mitigate this issue assume task labels (which are often not available in practice), employ brittle regularization heuristics, or use off-policy methods that suffer from instability and poor performance. We present Locally Constrained Policy Optimization (LCPO), an online RL approach that combats CF by anchoring policy outputs on old experiences while optimizing the return on current experiences. To perform this anchoring, LCPO locally constrains policy optimization using samples from experiences that lie outside of the current context distribution. We evaluate LCPO in Mujoco, classic control and computer systems environments with a variety of synthetic and real context traces, and find that it outperforms a variety of baselines in the non-stationary setting, while achieving results on-par with a "prescient" agent trained offline across all context traces. LCPO's source code is available at https://github.com/pouyahmdn/LCPO.

comment: ICLR '25 Spotlight

♻ ☆ Wasserstein multivariate auto-regressive models for modeling distributional time series

Yiye Jiang, Jérémie Bigot

This paper is focused on the statistical analysis of data consisting of a collection of multiple series of probability measures that are indexed by distinct time instants and supported over a bounded interval of the real line. By modeling these time-dependent probability measures as random objects in the Wasserstein space, we propose a new auto-regressive model for the statistical analysis of multivariate distributional time series. Using the theory of iterated random function systems, results on the existence, uniqueness and stationarity of the solution of such a model are provided. We also propose a consistent estimator for the auto-regressive coefficients of this model. Due to the simplex constraints that we impose on the model coefficients, the proposed estimator that is learned under these constraints, naturally has a sparse structure. The sparsity allows the application of the proposed model in learning a graph of temporal dependency from multivariate distributional time series. We explore the numerical performances of our estimation procedure using simulated data. To shed some light on the benefits of our approach for real data analysis, we also apply this methodology to two data sets, respectively made of observations from age distribution in different countries and those from the bike sharing network in Paris.