Action
-
AdaDexTrack: Dynamic Modulation for Adaptive and Generalizable Dexterous Manipulation Tracking
2026-10-24 18:46:10Jianibieke Adalibieke, Qianwei Han, Xueyi Liu, Yuzhe Qin, Li YiAbstract
Language is a natural way to command robots, but converting a single instruction into a long-horizon, contact-rich hand-object interaction remains challenging: synthesized references are noisy, human-to-robot retargeting introduces embodiment bias, and fixed-reference tracking lets small errors snowball. We address this with AdaDexTrack, a modulator-in-the-loop framework for language-conditioned manipulation tracking. A distilled generalist tracker serves as the skill carrier, while a tightly aligned modulator performs three feedback corrections: reference modulation (continual adjustment of what to track), object-latent modulation (online adaptation of the object representation to recruit suitable skills), and positional-target modulation (small state-dependent refinements for execution). The tracker is learned via large-scale specialist to generalist distillation on a corpus of language-conditioned hand-object trajectories; the modulator is trained with RL under the same task objective, ensuring tight coupling. Across large-scale evaluations, AdaDexTrack consistently outperforms prior SOTA on unseen-trajectory and unseen-object sets in both average tracking error and success rate, demonstrating robustness and generalization. We further show zero-shot sim-to-real transfer on real hardware, where adding the modulator yields substantial gains over a tracker-only variant. AdaDexTrack reframes language-conditioned dexterous manipulation as modulated tracking, replacing the open-loop, fixed-reference tracking with in-loop modulation that adjusts the reference, object latent, and positional target, yielding drift-resistant execution from noisy text references.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33852
PDF
-
EnvLunch: The Environment as a Free Lunch for Learning to Interact
2026-10-24 17:29:24Wei Liu, Min Gao, Yali Du, Yulan HeAbstract
For agentic data analysis, language-model agents often require a large number of annotated tasks to learn how to interact effectively with executable environments (i.e., databases). While high-quality annotated tasks require heavy manual effort, an executable environment already provides rich interaction structure and verifiable outcomes. In this sense, the environment itself offers a free lunch, an additional set of verifiable tasks that require no human annotation. We propose EnvLunch, a warm-up stage that turns this environmental signal into agent training without human-written or model-generated task descriptions, before downstream agentic reinforcement learning. EnvLunch samples valid action trajectories, executes them to obtain final observations, and trains an agent to recover a functionally equivalent trajectory given only its final observation. This label-free warm-up enables the model to outperform the baseline by over 9 points on average across all downstream training budgets, and with few labels, the baseline needs 8 to 32 times as many annotated examples to match the accuracy that the EnvLunch model reaches at its smallest training budget. Letting the environment alone define the task yields a larger gain than having the model write its own tasks over the same environments, and this warm-up also improves fine-tuning on benchmarks beyond data analysis. EnvLunch therefore saves annotated tasks and provides a better starting point for fine-tuning agentic models when such tasks are limited.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33817
PDF
-
IntentDelta: A Human-Constructed Benchmark for Evolving User Intent
Abstract
Language-model agents must adapt as users revise their requirements and priorities during complex tasks. We introduce IntentDelta, a human-constructed multi-turn benchmark with 638 trajectories and 3,766 turns across WebShop and TravelPlanner. Each turn pairs a user utterance with a cumulative intention state and human reference actions. The trajectories are cross-verified and cover feasibility- and preference-driven changes in constraints and priorities. We evaluate eleven agents on intention recovery and action quality. Average Intention Recovery is 33.1\% in TravelPlanner and 44.7\% in WebShop, while average Action Score is 31.6% and 41.0%, respectively. Agents struggle to preserve earlier requirements and distinguish soft preferences from must-have constraints. Correct intention recovery also does not guarantee successful execution, highlighting challenges in translating evolving user intentions into actions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33806
PDF
-
Are We Done with UMLS Injection into Language Models for Medical Question Answering?
2026-10-24 17:06:58Yiran Zeng, Giacomo Frisoni, Lorenzo Molfetta, Gianluca MoroAbstract
Maybe not. The Unified Medical Language System (UMLS) has become an oracle for millions of clinicians and health system developers worldwide. Researchers have increasingly used knowledge graphs derived from UMLS to enhance language models' factual accuracy and reasoning traceability in high-stakes tasks. Medical Question Answering (MQA) is a prime testbed for integration, with benchmarks usually involving professional exercises such as disease diagnosis and drug prescription from patient history. Despite the assumed benefits, empirical evidence shows that UMLS-based graph injection typically yields modest performance gains. This paper reviews existing MQA work, critically analyzing the selected UMLS sources and their usage with language models, and probes each of these choices in isolation. We uncover methodological issues never discussed in prior work, from low UMLS coverage to weak subgraph retrieval pipelines. Based on our findings, we provide actionable recommendations to unlock the untapped potential of UMLS as an external knowledge source.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33801
PDF
-
The Hitchhiker's Guide to System 1 Models: Fast Decision-Making Beyond LLMS
2026-10-24 16:26:01Haggai RoitmanAbstract
Large Language Models excel at reasoning and generation but their autoregressive, token-by-token inference makes them unnecessarily slow, expensive, and poorly calibrated for the many routine decisions (routing, gating, scoring, classification) that modern AI systems require at sub-second latency. A growing class of models addresses this gap: System 1 decision models, architectures that map an input directly to a typed, structured decision in a single forward pass. This guide provides a comprehensive, practitioner-oriented treatment of the field. We propose a five-family taxonomy covering encoder-based classifiers, reward models, non-autoregressive decision engines, contrastive decision models, and extraction models, and we provide in-depth coverage of each family's architecture, training methodology, and deployment trade-offs. A central theme is calibration: we develop the theory of proper scoring rules, analyze Reinforcement Learning for Calibrated Decisions (RLCD), and examine recently discovered phenomena including IIA violations, ordinal-scale bias, and few-shot recalibration under distribution shift. On the integration side, we describe how System 1 models serve as decision infrastructure within agentic workflows and System 1 + System 2 cascades that route between fast and slow models based on calibrated confidence. The guide covers 18 application domains, maps the rapidly growing open-source ecosystem (over 70 model reproductions in weeks), provides runnable code throughout, and includes a practitioner's checklist, an open-source model zoo, and 22 self-assessment questions with detailed answers. Spanning 190 references, it is intended as a unified entry point for researchers and engineers working at the intersection of fast decision-making, calibration, and production AI systems.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33775
PDF
-
CFL-aware edge design for graph neural solvers of hyperbolic transport
2026-10-24 16:20:20Jinyoung Jeong, Joseph B. Choi, Xinlun Cheng, H. S. Udaykumar, Sanghun Choi, Stephen S. BaekAbstract
Hyperbolic partial differential equations (PDEs) transport information at finite speed along characteristics, often producing shocks and discontinuities that are challenging to capture numerically. Graph neural networks (GNNs) offer learned update rules, but neighbor aggregation alone does not tell a GNN how far information must travel and along which edge. As a result, when a GNN model takes a time step larger than what was used in training, shock fronts smear and oscillations grow. In this work, we ask whether making the scale of transport explicit as an edge-local Courant-Friedrichs-Lewy (CFL) signal can enhance GNNs' ability to model hyperbolic PDEs. Holding the architecture fixed, we vary only the information given to each edge, from a raw edge form to forms that progressively expose the edge-local transport interaction through direction, length, and wave speed. Because all of these are computed from the same ingredients, what changes is not the amount of information but its form. The results show a consistent pattern, in which feature count is not the limiting factor, and appending CFL directly to a raw edge form produces almost no benefit. The CFL conditioning proves most effective when the edge information already exposes the transport interaction, reducing final-step error by 10.4-14.4%, and the improvement widens as the step size grows larger. Notably, this improvement arises not from smearing the solution but from suppressing oscillations near shocks. These results point to a specific design principle that CFL conditioning is most effective when the edge representation already separates the direction of interaction from spatial scale, so that the CFL value has a well-defined quantity to modulate.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33769
PDF
-
Computer Vision Navigation for an Autonomous Vehicle Model Using Reinforcement Learning
2026-10-24 15:52:59Dinis Fernandes, Tiago Ribeiro, Ant\'onio Ribeiro, Carolina Lopes, Gil Lopes, A. Fernando RibeiroAbstract
This work aims to assess the capabilities of the reinforcement learning algorithm, Proximal Policy Optimization (PPO), in developing a policy capable of controlling the speed and steering angle of a vehicle model to navigate within simulated road lines. The control system's input is a camera, and its outputs are the rotational speed of the rear wheels and the steering angle of the front wheels, controlling the vehicle's speed and direction, respectively. The algorithm is trained and tested in a simulated environment, where it receives image frames from the vehicle's camera, processes them through a feature extraction algorithm, and feeds them into the policy. For each frame, the policy outputs speed and direction commands for the vehicle. The policy is iteratively updated to enhance its performance. The reinforcement learning algorithm successfully generates a policy capable of navigating the vehicle in a simplified environment. Potential improvements are suggested to enhance the system's robustness in more complex scenarios. Creating a simulation environment suitable for testing an advanced reinforcement learning algorithm is a complex task. This project offers a framework for other developers to test different approaches and algorithms. The tested algorithm, PPO, uses proximity intervals - an innovative approach in reinforcement learning. In some scenarios, the algorithm achieved better performance than predecessor algorithms, which are typically harder to implement and require tuning more hyperparameters.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33755
PDF
-
ProtAgent: Post-Training Compact Protein Tool Agents with Interaction-Level Credit Assignment
2026-10-24 15:42:13Yuxiang Wang, Baojian Zhou, Qifang Zhao, Keyue Jiang, Yanghua Xiao, Xiaoxiao XuAbstract
Protein annotation agents must choose suitable bioinformatics tools, interpret their outputs, and use the resulting evidence in the final answer. Most existing systems rely on frontier-scale models. We ask whether a compact model can instead learn this behavior through post-training with real tool interaction. We present ProtAgent, a two-stage pipeline that uses supervised fine-tuning (SFT) to learn the tool interface and standard on-policy REINFORCE to improve the interaction policy. Intermediate tool calls receive action rewards, and the final response receives a terminal answer reward. Step-level returns assign these rewards to the responses that produced them and to preceding actions. The episode-level baseline instead assigns one episode return to every action. On a fixed closed-loop protocol with real bioinformatics tools, ProtAgent-8B achieves 61.00% success under the annotation-coverage metric, compared with 41.60% for DeepSeek-V3.2-Thinking 671B and 60.80% for the strongest evaluated frontier system. In the three-seed comparison, mean success is 42.70% with episode-level returns and 61.40% with step-level returns under identity return normalization; the corresponding group-relative results are 36.50% and 58.20%. The metric measures recovery of reference annotations under this protocol; it does not by itself establish biological correctness or complete evidence grounding.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33746
PDF
-
GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation
2026-10-24 15:41:00Renhang Liu, Wenzhi Zhao, Zhuo Yang, Liliang Chen, Pengfei Zhou, Shengcong Chen, Guanghui Ren, Youlun Peng, Rongjun Jin, Nan Wang, Sukai Wang, Xindong He, Jinyuan Feng, Ziyu Xiong, Linqing Zhong, Yifei Wei, Feng Han, Long Zhang, Da Huang, Nanshu Zhao, Chenghao Yin, Mo Wu, Zhaodong Yan, Kongtao Hu, Yuxiang Yan, Aogelijiang Niyazi, Yu Fang, Jia Zeng, Lizhu Meng, Daizhen Lv, Haoyu Cao, Zhiwen Hou, Lianjin Ye, Yuehan Niu, Zhikai Cai, Xuan Hu, Hui Min, Xiongfeng Cai, Yue Liao, Jing Wu, Soujanya Poria, Ye Li, Sanping Zhou, Maoqing Yao, AgiBot Research TeamAbstract
World-action models (WAM) predict future states to guide robot actions, enabling learning from both action-free video and action-labeled interaction. Most inherit pretrained video generators, leaving WAM pretraining and scaling underexplored. We introduce Genie Envisioner Act 2.0 (GE-Act 2.0), a world-action model whose trainable generative and action components are all initialized from scratch on manipulation data. It combines a control-oriented autoencoder (CoAE), a single-step visual planner (SVP), and an inverse dynamics model (IDM). CoAE retains action- and instruction-relevant information under aggressive compression, while SVP produces a complete future state in one differentiable pass, so visual planning and inverse dynamics can be pretrained separately on complementary data. The components are then jointly trained with knowledge-aligned selective optimization (KASO), which reduces mismatched supervision by selecting only predicted futures judged behaviorally compatible with the recorded action. We evaluate pretrained checkpoints directly, without per-task fine-tuning, on 100 tasks across 20 manipulation skill groups with held-out scenes, backgrounds, lighting, and object instances. Scaling co-training data from 300 to 30,000 hours raises success from 17.1% to 44.1% on G1-OP and from 13.4% to 31.1% on G2-90D; despite comprising less than 2% of the co-training data, G2-90D improves by 17.7 points, suggesting cross-embodiment transfer. Gains span 19/20 and 18/20 skill groups, and skill-specific coverage strongly correlates with zero-shot out-of-distribution (OOD) success (Pearson r=0.80; Spearman rho=0.85). Under the same protocol, the model grounds object, color, shape, and position references in at least 90% of trials and follows explicit instructions even when they conflict with an already-committed behavior or a conventional scene association.
Abstract (translated)
URL
https://arxiv.org/abs/2609.05588
PDF
-
Floor Friction Assessment from Robot Tire-Surface Interaction
2026-10-24 15:24:25Kulbhushan Komarraju, Nathaniel Pearson, Yu Gu, Jason GrossAbstract
Floor traction influences wheel-slip behavior, dead-reckoning accuracy, and the navigation performance of wheeled mobile robots that rely on wheel odometry. As these robots become increasingly common in GNSS-denied indoor environments, including automated manufacturing and retail facilities, their onboard measurements could also be used to characterize floor conditions and improve localization. In this study, a skid-steer robot was driven over five floor surfaces with distinct traction characteristics using commanded velocity profiles designed to induce wheel slip through repeated acceleration and direction reversal. Differences between the body velocity estimated from inertial sensing and the velocity obtained from wheel odometry were evaluated using the integrated relative-velocity error, phase-derived wheel-to-body response delay, and velocity-reversal response delay. Welch power spectral density analysis was additionally applied to characterize the frequency distribution of the velocity mismatch during positive and negative acceleration. Band-integrated spectral power, spectral centroid, and spectral decay rate were extracted from each recording and evaluated using linear discriminant analysis and multiclass classification. The response-delay measurements produced consistent surface-dependent trends, with the longest delays observed on soap and the shortest delays observed on carpet. The spectral features also provided clear separation among the tested surfaces, achieving a balanced classification accuracy of 97.1% under recording-level cross-validation. These results demonstrate that both the magnitude and frequency domain characteristics of wheel-to-body velocity mismatch contain useful information about tire-surface interactions. Such information could support surface recognition and the development of slip-aware wheel-odometry weighting within robot state-estimation frameworks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33735
PDF
-
Human-Feedback-Guided Multi-Agent Inverse Reinforcement Learning for Interactive Driving Behavior Modeling
2026-10-24 15:14:41Longhao Yan, Jinhao Liang, Kaidi YangAbstract
Modeling human driving behavior is fundamental to understanding driver decision-making and supporting autonomous vehicles (AVs) in mixed traffic. Reward learning offers a principled way to infer the latent objectives behind driving behaviors, yet most existing methods are formulated in single-agent settings and treat surrounding vehicles as dynamic obstacles. Directly extending reward learning to multi-agent scenarios is difficult because the joint action space grows exponentially with the number of interacting vehicles. In addition, existing approaches mainly rely on physical demonstrations and rarely exploit human feedback during reward recovery. To address these limitations, this paper proposes a Human-Feedback-Guided Multi-Agent Inverse Reinforcement Learning (HFG-MAIRL) framework for interactive driving behavior modeling. First, we develop a multi-agent maximum-entropy inverse reinforcement learning module that explicitly models each vehicle and its directly interacting neighbors in a localized Markov game, while using aggregated traffic states to capture the influence of the broader traffic environment. The resulting coupled multi-vehicle responses are approximated through finite marginal-response iterations over factorized vehicle-level policies, avoiding explicit enumeration of the full joint action space. Second, we incorporate human feedback as ordinal preference supervision through a Bradley-Terry likelihood. In our experiments, driver emotion labels are used as one instantiation of such feedback and converted into pairwise ordinal preference constraints. Experiments on the AIDE dataset show that HFG-MAIRL outperforms representative IRL baselines in decision modeling and produces interpretable feedback-aligned reward patterns.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33726
PDF
-
Traffic-Context-Aware Driver Emotion Recognition with Expectation-Deviation-Driven Inference
2026-10-24 14:59:21Longhao Yan, Jinhao Liang, Kaidi YangAbstract
Driver emotion provides an important indication of how drivers perceive and evaluate ongoing traffic situations, making reliable emotion recognition an important capability for effective human-vehicle interaction. Yet, most existing driver emotion recognition approaches rely primarily on in-cabin images and overlook how surrounding traffic interactions shape driver emotional responses. To address this limitation, we propose a traffic-context-aware driver emotion recognition (TADER) framework that integrates in-cabin signals with out-of-cabin traffic scenarios, modeling driver emotions as arising from expectation deviation represented by the discrepancy between predicted and observed vehicle trajectories. The proposed framework comprises two key components: (1) a vision-based base recognizer that extracts spatiotemporal representations from in-cabin facial images to infer driver emotions, and (2) a traffic-context-aware fusion module that quantifies expectation deviation from external traffic scenarios and integrates it into the base recognizer via attention-based embedding fusion. To overcome the scarcity of paired in-cabin and out-of-cabin data, we introduce a transfer learning strategy that transfers a trajectory prediction model trained on large-scale naturalistic datasets to smaller paired datasets. Extensive experiments demonstrate that TADER consistently outperforms state-of-the-art methods across both naturalistic and simulated driving environments by incorporating traffic context information into emotion recognition, resulting in improvements of 5% in accuracy on the AIDE and manD datasets.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33713
PDF
-
Open-World Law-to-Action: Vision-Language Priors and Touch for Parts with Hidden Physics
2026-10-24 14:41:14Mohamed A. MabrokAbstract
A robot in a home it has never seen must open doors, drawers, lids and taps whose physics a picture does not show. Instructions such as "open the drawer halfway, gently" depend on that physics: how far a part travels, how it resists, and what locks it. We present Open-World Law-to-Action (L2A). A vision-language model (VLM) proposes, for each movable part, its motions, the terms of its law of motion, and what may lock it. Each proposal becomes a prior with a floor, so that touch can always overturn it. The robot identifies each part's law online by Bayesian linear regression with closed-form evidence, for parts that move in one way or in two, and keeps a hypothesis that an unseen part or motion locks the target. We prove that touch alone cannot rule this hypothesis out, derive when another look is worth its cost, and show that a constant load and friction cannot be told apart while a part moves one way. All experiments are in simulation, developed on some scenes and run once on held-out ones. In new kitchens the robot completes 80% of tasks from the VLM's proposals, against 98% when every part is known. Proposed physics raises success on objects that need new terms from 353 to 378-383 of 400 episodes, and a second look raises success in mixed scenes from 284 to 329 of 366 tasks. A mobile robot that takes nothing from the simulator completes 70 of 100 requests in new scenes, against 62 when the simulator says which part each box is (after one change made between two held-out tests).
Abstract (translated)
URL
https://arxiv.org/abs/2610.33705
PDF
-
Gender Associations in LLM-Mediated ADHD Self-Diagnosis
2026-10-24 14:24:11Matilde Elene Hansen, Ross Deans Kristensen-McLachlanAbstract
Large Language Models (LLMs) are increasingly used for medical self-diagnosis. However, the way demographic cues shape outputs in these interactions remains poorly understood. This is particularly consequential for diagnoses like ADHD, where persistent gender disparities are not fully aligned with the clinical evidence. We adapt the Context Association Test (CAT) to evaluate whether GPT-4o responds to gendered ADHD associations in a zero-shot, self-diagnostic setting. We show that explicit cueing consistently amplifies certain ADHD subtype associations while implicit cueing produces near proxy-aligned distributions. These divergences suggest that explicit demographic cues may activate bias-mitigation mechanisms that over-correct rather than eliminate bias, whereas implicit cues appear to reflect the model's default behavior. Our results demonstrate that evaluating gendered bias in LLM-mediated self-diagnosis requires careful attention to the demographic associations a model encodes.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33697
PDF
-
EXE: Egocentric Cross-Embodiment Pretraining via Factorized Latent Actions for VLA Models
2026-10-24 14:08:42Qingyuan Cai, Bo Zhang, Yongzhen Huang, Jianyu ChenAbstract
Vision-Language-Action (VLA) models are bottlenecked by the scarcity of robot data. Egocentric human manipulation videos offer an abundant, diverse, and semantically-rich alternative that inherently captures manipulation intent. However, the embodiment gap makes it difficult to extract action-relevant representations from these videos. From the perspective of action execution, human hands and robots share analogous manipulation primitives---reaching, grasping, placing---yet exhibit distinct kinematics tied to their embodiments. This might motivate our key approach: egocentric video pretraining can yield shared action representations capturing transferable manipulation skills, while human-specific motion patterns should be disentangled to avoid interfering with robot policy learning. To this end, we propose EXE, an Egocentric Cross(X)-Embodiment pretraining framework that learns a factorized latent action representation from egocentric human videos, separating shared manipulation priors from embodiment-specific signals. By leveraging the shared action space, the robot policy acquires generalizable manipulation semantics from diverse human demonstrations without being biased by human-specific kinematics. This shared knowledge transfers across embodiments with different kinematic structures, enabling the same pretraining to benefit both parallel grippers and dexterous hands. EXE yields consistent gains over baseline without pretraining on RoboTwin 2.0, RoboCasa GR1, and real-world manipulation tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33685
PDF
-
Extraction meets Distillation - cryptanalytic techniques to recover parameters from a Deep Neural Network
2026-10-24 13:36:49Biprarshi Biswas, Arya Anshuman Ojha, Yubaraj Das, Devansh Khosla, Kamran Rais, Saankhya Samanta, Subhamoy Maitra, Pantelimon StanicaAbstract
We study fully connected neural-network reconstruction by combining cryptanalytic weight-direction extraction with learning from predicted labels. We adapt a ReLU extraction implementation to Leaky ReLU, compare sign-search methods, and estimate biases and output layers. We prove conditional bounds for direction recovery with approximate prefixes, preservation of local distance votes, prediction changes from aligned parameter errors, and sign selection with independent scoring data. We also specify exact row restoration and distinguish its guarantee from that of gradient masking. The experiments evaluate assisted extraction, with privileged access to gradients, hidden activations, and victim parameters, followed by label-based reconstruction. Across eight search seeds on the largest synthetic architecture, Leaky ReLU recovers 51--57 neurons in the deepest hidden layer, while ReLU recovers none. This advantage does not hold uniformly across architectures and layers. Removing victim-weight access from the sign walk gives less accurate signs, despite high final prediction agreement. Across 40 seeded runs on the two smallest synthetic victims, agreement on 5,000 inputs excluded from fitting and selection ranges from $0.9890$ to $0.9992$. On two CIFAR-10 victims, the reconstructed models achieve $56.40\%$ and $58.29\%$ agreement and exceed the distillation baseline on three functional metrics. We report label requests separately from privileged access and explain why prediction agreement does not identify network parameters.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33671
PDF
-
SRHarness: A Harness for Agentic Symbolic Regression
2026-10-24 13:20:10Zihan Yu, Shixuan Zhou, Hao Huang, Jingtao Ding, Yong LiAbstract
Recent agentic symbolic regression approaches increasingly rely on large language models to analyze data, select scientific operations, and refine hypotheses over long search trajectories. In such systems, performance depends not only on the underlying model and search strategy, but also on the runtime infrastructure that supports scientific search. We introduce SRHarness, a domain-specific harness for agentic symbolic regression built around three mechanisms: composable scientific actions that provide a common interface over raw, transformed, and candidate-derived quantities; persistent scientific state that retains evaluated hypotheses and exposes compact model-facing views; and trajectory lifecycle management that coordinates continuation, branching, restart, and termination. On LLM-SRBench, SRHarness consistently improves both numerical generalization and symbolic recovery under matched LLM backbones. With DeepSeek-v4-flash-0731, it achieves 93.69% symbolic accuracy on LSR-Transform, compared with 62.16% for SR-Scientist, and retains 72.97% accuracy on an anonymized variant that removes scientific descriptions and variable semantics, versus 39.64% for SR-Scientist. Under the same DeepSeek-v4-flash-0731 backbone, SRHarness also substantially outperforms Codex (72.97% vs. 20.72%) and reaches performance comparable to Codex with GPT-5.5, while simply providing Codex with the same scientific tools does not reproduce this advantage. These results show that effective agentic symbolic regression depends not only on models or tools, but also on structured runtime support for organizing scientific actions, accumulated hypotheses, and long-horizon search.
Abstract (translated)
URL
https://arxiv.org/abs/2609.35501
PDF
-
Power-of-Two Leaky Integrate-and-Fire Neurons for Event-Driven Vision-Language-Action Inference
2026-10-24 13:16:20Rui Liu, He Lv, Ziyao Zhang, Rui Ma, Rongxiang NieAbstract
Vision-language-action (VLA) models have demonstrated strong capabilities in robotic manipulation, yet their reliance on dense, continuous-valued computation presents a fundamental challenge for energy-constrained deployment. Inspired by the sparse, event-driven signaling and temporal integration mechanisms of biological neurons, a promising alternative is to introduce spiking computation into VLA models. However, enabling such computation without restructuring pretrained policies or compromising closed-loop control remains an open challenge. To address this challenge, we introduce the Power-of-Two Leaky Integrate-and-Fire (PoT-LIF) neuron, which emits signed spikes with power-of-two amplitudes. By combining magnitude-coded spike events with residual-preserving membrane dynamics, PoT-LIF enables continuous information to be progressively represented through discrete spikes. The Heartbeat formulation further localizes this process within weighted projections, bridging biologically inspired spiking computation and pretrained VLA architectures without modifying their external inference interfaces. We integrate PoT-LIF into the vision projection, vision-language projection, and action projection of $\pi_{0.5}$ and evaluate both operating modes over 2,000 paired closed-loop LIBERO episodes per configuration. Standard-mode PoT-LIF VLA achieves success rates of 96.90% and 96.35% at $T = 1$ and $T = 2$, respectively, while Heartbeat mode achieves 97.25% and 96.90% at $T = 2$ and $T = 4$, compared with 96.70% for the original continuous-activation $\pi_{0.5}$ baseline. At $T = 1$ with a 60.01% firing rate, PoT-LIF uses approximately 1/7.67 of the dense-reference arithmetic energy, corresponding to a modeled reduction of 86.95%.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33658
PDF
-
Physics-Informed Graph-Based Safe RL with Constraint-Specific Critics for Community Battery Coordination in Distribution Network
2026-10-24 13:11:29Chenghao Huang, Yue Yang, Hao WangAbstract
High penetrations of distributed energy resources (DERs) and fast-varying electric vehicle (EV) charging intensify voltage and line-limit violations in distribution networks (DNs), making safe, low-cost operation increasingly challenging. This paper proposes a physics-informed safe deep reinforcement learning (DRL) framework for coordinating community battery energy storage systems to mitigate constraint violations while minimizing operating cost. To represent nonlocal spatiotemporal interactions, we develop a dual-level physics-informed graph representation in which line impedance and capacity shape spatial information propagation, while power-balance, voltage-profile, and flow-reconstruction losses regularize the latent state toward control-relevant physical consistency. Building on this representation, we develop Lagrangian Dedicated-Critic PPO (LDCPPO), which implements constraint-specific value decomposition: independent reward and cost critics preserve heterogeneous voltage- and line-loading-related learning signals and combine their advantages only during the Lagrangian policy update. Case studies on DNs with high rooftop photovoltaic penetration and EV charging demonstrate that the proposed method reduces the key voltage-violation degree by up to 80.5% compared with the tested baselines, while also achieving the lowest voltage-violation count and line-overload degree with competitive operating cost.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33654
PDF
-
TAD: Distilling Vision-Language-Action Policies from Teacher Trajectories Alone
2026-10-24 13:05:18Yun Zhu, Da WangAbstract
Online distillation of vision-language-action (VLA) policies commonly relies on teacher action scores that an inference-only interface may not expose. Trajectory-only Adversarial Distillation (\gad{}) instead turns a fixed bank of teacher joint-position features into adaptive feedback: a discriminator compares feature marginals from teacher and student rollouts and supplies rewards for on-policy learning. The student interacts with the environment without online teacher queries or environment task rewards. On ManiSkill \texttt{PutOnPlateInScene25Main}, \gad{} improves OpenVLA-OFT (7B) from $26.2\%$ to $76.6\%$ \successonce{} ($+50.4$ percentage points) and from $20.3\%$ to $58.6\%$ \successend{}. It trails white-box on-policy distillation (OPD) at $81.6\%$ / $59.4\%$ by $5.0$ / $0.8$ percentage points ($n{=}256$ evaluation trajectories per checkpoint). The tested four-epoch continuation achieves the highest \gad{} result without additional teacher data. This best run uses $5\times$ OPD's student interaction budget and $12.5\times$ its complete-rollout batch-passes. The comparison demonstrates online VLA improvement without teacher scoring while quantifying unequal interaction and optimization budgets.
Abstract (translated)
URL
https://arxiv.org/abs/2610.33650
PDF
