BEAKER: An Expert-Curated Benchmark for Embodied Brains in Self-Driving Chemical Laboratories
Fei Lin, Tengchao Zhang, Ziyang Gong, Xiaotong Yu, Bohan Zhang, Yifan Zhou, Qihao Yang, Ji Dai 等 18 位作者
AI 速览：作者提出并构建了专家人工标注的 BEAKER 基准，用于评测多模态大模型在自驱动化学实验室中的执行导向具身能力。
Self-Driving Chemical Laboratories (SDCLs) are moving chemical experimentation toward embodied automation, where machines must perceive laboratory scenes, reason about experimental states, and act under physical, procedural, and safety constraints. Although Multimodal Large Language Models (MLLMs) are increasingly considered as embodied brains for such systems, existing laboratory benchmarks mainly focus on safety, anomaly detection, or general scientific understanding, leaving their execution-oriented embodied capabilities underexplored. To fill this gap, we introduce BEAKER, a Benchmark for evaluating Embodied Actionability, Knowledge, Experimentation, and Reasoning in SDCLs. BEAKER contains 1,000 expert-curated question-answer samples across 11 image and video task types, covering key judgments required for chemical laboratory execution, including affordance grounding, trajectory planning, equipment, state, and safety recognition, apparatus setup, operation and step reasoning, and temporal procedure reasoning. All questions and annotations are manually created and reviewed by chemistry and embodied Artificial Intelligence (AI) experts, without AI-assisted generation or annotation. We evaluate 28 general-purpose, science-oriented, chemistry-specialized, and embodied MLLMs in an offline, single-turn setting. Results show that current models can follow coarse experimental procedures in videos, but still struggle to ground actions into operable regions and feasible trajectories, to recognize fine-grained laboratory apparatus, states, and safety norms, and to order steps in mid-length procedures. BEAKER thus reveals a clear gap between coarse procedural understanding and the precise, action-grounded perception required for laboratory execution, providing a diagnostic benchmark for future MLLMs in SDCLs.
Tool-using large language model (LLM) agents translate model outputs into consequential actions, turning indirect prompt injection into an execution-security problem. Existing defenses constrain tool calls, trajectories, or information flows, but often leave action semantics, state dependencies, and evidence relations implicit, limiting generalization across heterogeneous tools and complicating safe updates. We present \ours{}, a dynamic ontology-guided governance framework. \ours{} grounds heterogeneous calls in a shared semantic representation and combines an executable ontology with episode-level task state to derive deterministic ready, waiting, and forbidden decisions. A waiting decision identifies an unmet but recoverable obligation, enabling bounded replanning without expanding the agent's authorization boundary. \ours{} evolves its ontology offline from verified training executions. An offline LLM proposes instance-independent action--state axioms, which are accepted only if they address ontology errors exposed by training failures, preserve logical consistency and protected entailments, and pass held-out validation before the ontology is frozen. We evaluate \ours{} on AgentDojo, ASB, and AgentDyn under benign and indirect prompt-injection settings. \ours{} achieves a 0\% attack success rate (ASR) across all three benchmarks while improving task utility under attack compared with existing baselines. These results show that explicit semantic governance can strengthen execution security without indiscriminate blocking.
Privacy-preserving continual learning (PPCL) must reduce the reproduction of sensitive content while retaining useful knowledge across sequential tasks. Formal privacy guarantees characterize randomized mechanisms, whereas operational output control concerns whether a trained model selectively reduces the likelihood of sensitive content in its outputs. In this work, we investigate the latter together with continual-learning utility under realistic task evolution. Retention and privacy correction operate at different granularities: task acquisition requires broad preservation of current- and old-task behavior, whereas privacy correction targets sparse annotated positions. Joint optimization leaves the current-task preservation target continually changing. We propose SPARK, a retention-correction decomposition that first freezes the learned post-task distribution and then applies selective correction around this stable reference. Self-Distillation Replay learns the current task while distilling behavior from previous tasks, and Post-Task Privacy Correction reduces annotated-PII likelihood while anchoring current- and old-task non-PII behavior to the resulting checkpoint. Extensive evaluations demonstrate that SPARK achieves effective selective PII suppression while preserving strong continual-learning utility and knowledge retention across diverse settings. Code and data will be released upon publication.
Vision-language verifiers judge whether an image faithfully matches a textual description, yet existing training pipelines typically rely on static supervision from stronger external judges or curated hard-negative data. Such supervision requires external generation or annotation, is teacher-bounded, and is difficult to adapt as model failures evolve. We introduce PROVE (Perturb-and-Repair Optimization for Vision-Language Evaluation), a self-play framework that trains verifiers through a closed perturb-verify-repair-verify loop. Starting from a known-correct image-caption pair, a Mutator introduces a typed perturbation over objects, attributes, counts, or relations. The verifier then judges the perturbed caption and provides textual feedback, which a Repairer uses to recover a faithful caption for re-verification. These two verification rounds produce structured rewards for informative mismatches and visually grounded repairs. Since perturbation, verification, and repair are role-conditioned behaviors of the same evolving multimodal backbone, updating the shared parameters also improves the verifier itself. Experiments show that PROVE improves fine-grained compositional verification across model scales and families and strengthens visual-detail robustness on the reference backbone, with limited, benchmark-dependent changes on broader multimodal tasks.
Jun Huang, Junyang Zhang, Zijie Yue, Bingkun Wang, Yong Luo, Miaojing Shi
Video Reasoning Segmentation (VRS) aims to infer and segment target objects specified by reasoning queries in videos. Existing methods typically first infer the target and generate a few segmentation tokens for it, which are then fed into a mask decoder for mask prediction. However, such segmentation tokens tend to be both semantically monotonous and spatially ambiguous, making them largely insufficient to represent the target to guide mask prediction. In this paper, we propose an explicit visual object grounding framework for VRS. Specifically, we first propose a spatial-aware prompt generation and refinement scheme, in which we query a multimodal large language model to generate explicit spatial prompts (i.e., bounding box and point set) for the target and then refine them through multi-step dialogue. Furthermore, we introduce a query diversification and alignment module to generate auxiliary queries that describe the same target from different perspectives. We then enforce cross-query consistency between their predicted spatial prompts to alleviate the training bias caused by query limitations. Finally, we select several reliable frames using a prompt-based keyframe selection strategy to support mask decoding and propagation. Extensive experiments on eight datasets demonstrate that our method significantly outperforms state-of-the-art methods. The code is available at https://github.com/viscom-tongji/EVOG.
Large language model-based agents increasingly rely on external skills to reuse task-specific knowledge without retraining the model. As skills become common for extending agent capabilities, understanding creation and evolution is critical for reliable agent systems. However, existing studies mainly examine agent skills from a static perspective, leaving their evolution in real-world repositories largely unexplored. This limits our understanding of the patterns and motivations behind skill changes. To fill this gap, we present a large-scale empirical study of agent skills, capturing 84,200 skills and 750,898 commits from GitHub. We manually construct taxonomies of skill components, change patterns, and evolution motivations, and build an automated annotation pipeline to label changes at scale. Based on the taxonomy, we conduct quantitative analyses of skill composition and post-creation evolution frequency, together with qualitative assessments of modified components, change patterns, and motivations. Our findings reveal five implications: (i) reliability receives less attention in skill composition and change, motivating execution-contract skill design that incorporates verification and recovery mechanisms; (ii) evolution changes concentrate in project-coupled skills, calling for automated just-in-time skill updating mechanisms; (iii) change patterns are dominated by incremental enrichment and refinement, motivating pattern-guided skill enhancers that diagnose and improve incomplete skill components; (iv) redundancy is a practical maintenance issue, highlighting the need for context-efficient skill representation; (v) security-motivated changes show limited risk reduction, suggesting skill regression testing methods that objectively verify skill functionality and security. Beyond these insights, our dataset provides a basis for future studies on automated skill techniques.
Federated learning (FL) is well suited to Internet of Things (IoT) sensing. However, devices carry different sensors and come online in periodic patterns, and as such, the sensors among the participants change from round to round. Under naive averaging, devices without a sensor overwrite its parameters, and a single-sensor device never learns how its sensor relates to the others. Existing methods either redesign local training for mismatched sensors (assuming always-available devices) or select clients without regard to their sensors. We propose Tier-aware Intra-pod Decoupled Encoders (TIDE). Specifically, each sensor has its own encoder and auxiliary head, whose parameters are merged only within the "pod" (i.e., the devices that hold the sensor), while the fusion layer and the task head remain shared by all devices. On three multimodal activity recognition datasets against eight baselines, TIDE reaches 78.06% on ActionSense, 7.63 points above the next method, is the most accurate on all three datasets at half its own communication, and at equal communication leads on ActionSense and is ahead within seed variation on the other two.
Yuancheng Wei, Haojie Zhang, Linli Yao, Lei Li, Jiali Chen, Tao Huang, Yiting Lu, Duojun Huang 等 10 位作者
Image Difference Captioning (IDC) generates natural language descriptions that precisely identify differences between two images, serving as a key benchmark for fine-grained change perception, cross-modal reasoning, and image editing data construction. However, existing benchmarks lack diversity and compositional complexity, and standard lexical-overlap metrics (e.g., BLEU, METEOR) fail to capture semantic consistency or penalize hallucinations, which together prevent a comprehensive and robust evaluation of multimodal large language models (MLLMs) on IDC. To address these gaps, we introduce DiffCap-Bench, a comprehensive IDC benchmark covering ten distinct difference categories to ensure diversity and compositional complexity. Furthermore, we propose an LLM-as-a-Judge evaluation protocol grounded in human-validated Difference Lists, enabling a robust assessment of models' ability to both capture and describe visual changes. Through extensive evaluation of state-of-the-art MLLMs, we reveal significant performance gaps between proprietary and open-source models, highlight the critical importance of reasoning capability, and identify clear limitations in model scaling. Our framework also demonstrates strong alignment with human expert judgments and strong correlation with downstream image editing data construction quality. These findings establish DiffCap-Bench as both a reliable IDC evaluation framework and a practical predictor of downstream utility. The benchmark and code will be made publicly available to support further research.
Temporal Knowledge Graph Question Answering (TKGQA) requires language model agents to retrieve evidence, select relevant facts, and perform multi-step temporal reasoning. However, the trajectory-level outcome signal in GRPO does not explicitly distinguish the quality of intermediate decisions: downstream failures can penalize useful evidence acquisition, while identical rewards in all-failed groups eliminate the relative outcome signal despite differences in intermediate progress. We propose Credit Redistribution with Evidence-Driven Opportunities (CREDO), which uses intermediate evidence to recover local supervision, including in all-failed groups, and determine where outcome supervision acts. CREDO assigns positive stage-level credit through comparisons among rollouts of the same question and uses retained answer overlap as an opportunity proxy to shift part of the outcome-policy weight toward reasoning and answering. The framework preserves terminal rewards and trajectory-level advantages, while redistribution conserves total outcome-policy weight per trajectory. Following the public evaluation protocol of prior work, CREDO raises Overall accuracy from 78.0% to 82.4% and Multiple accuracy from 55.0% to 65.6% on MultiTQ relative to the strongest published baseline in our comparison. On TimelineKGQA, it raises cross-KG accuracy from 30.9% to 36.4% relative to our reproduced baseline via direct transfer to ICEWS Actor under matched training and evaluation settings.
Qin-Wen Luo, Sheng Ren, Xiang Chen, Rui Liu, Jun Fang, Naiqiang Tan, Jun Wang, Hua Chai 等 9 位作者
Deep search agents extend language models by interleaving reasoning with external retrieval. However, effective search-agent training requires addressing two challenges beyond final-answer optimization. First, the capability frontier of agents is largely determined by training questions, whereas fixed or labeled task collections cannot adapt as agents improve. Second, outcome-only reinforcement learning treats all successful trajectories equally, even when some correct answers rely on redundant searches, creating an inherent tension between answer accuracy and tool efficiency. To address these issues, we introduce \textsc{ZEST-Search}, a zero-label and tool-efficient framework for search-agent training via heterogeneous co-evolution. Starting from unlabeled reference questions, \textsc{ZEST-Search} constructs validated search tasks online through an answer-first and evidence-path-guided generation process. A heterogeneous proposer adaptively constructs training tasks around the solver's evolving capability boundary, while the solver learns search-augmented reasoning through reinforcement learning. To jointly optimize task difficulty and search efficiency, \textsc{ZEST-Search} employs an intermediate-difficulty reward for the proposer and a task-local tool-aware reward for the solver, encouraging informative tasks and correct answers with fewer redundant searches. Experiments on seven general-domain and multi-hop question answering benchmarks demonstrate that \textsc{ZEST-Search} consistently improves search-agent performance while achieving higher tool efficiency without labeled search-training data or expert search trajectories. Further analyses show that heterogeneous generation, adaptive difficulty control, and tool-aware optimization enable more effective search-agent training.
Bingchao Wang, Yongcheng Jing, Shunyu Liu, Chengyu Wang, Jun Huang, Jie Yang, Wei Liu, Dacheng Tao
Mobile GUI agents are typically fine-tuned 7B-scale vision-language models (VLMs) that generate each action as text. However, these VLMs decode every action token by token, and even agents that score candidate actions keep a large language model or VLM in the loop, making them slow and memory-hungry. To address these issues, we propose GUI-JEPA, to our knowledge the first decoder-free GUI agent policy built on a Joint-Embedding Predictive Architecture (JEPA), which selects actions in the latent space of a JEPA encoder pretrained on unlabelled GUI screenshots. Its action- and region-conditioned latent predictor turns the screen latent into one latent per candidate action, letting lightweight heads score all candidates against the instruction in a single batched pass, without decoding a token. We further show that the scorer relies on this latent pathway and that the predictor is trained for the decision itself: a randomly initialised predictor does just as well as one pretrained on screen dynamics. On AndroidControl, GUI-JEPA is 2.8-4.0x faster on the same GPU than a Qwen2.5-VL-7B agent we fine-tune with the same inputs, while reaching 86-92% of its in-distribution accuracy with 6% of its parameters. Code is available at https://github.com/bcwang-sjtu/GUI-JEPA
Bingchao Wang, Yongcheng Jing, Chengyu Wang, Jun Huang, Jie Yang, Wei Liu, Dacheng Tao
GUI grounding on high-resolution screens fails mostly on small targets, and zooming in is a common training-free remedy. However, most existing zooming pipelines do not match the zoom to the target's size, accept the zoomed answer unconditionally and never revisit a wrong answer. To address these issues, we propose AdaGround, a training-free framework with three components: (1) Token-Normalized Adaptive Zooming, which sets the magnification, field of view and zoom depth of each sample from the model's own estimate of the target size; (2) Verifier-in-the-Loop Refinement, which uses one marked-click Yes/No verifier to keep or reject each zoom, choose between zoom chains, trigger re-grounding and, with an existence query, decide when to abstain; and (3) Compositional Structure Grounding, which composes spreadsheet cells from their headers. On Qwen3-VL models, AdaGround surpasses state-of-the-art training-free zooming methods and many trained GUI models, improving accuracy over the backbone by 7.4 to 20.6 points on OSWorld-G and ScreenSpot-Pro, and its gains carry over to held-out UI-Vision. Code is available at https://github.com/bcwang-sjtu/AdaGround
Yifan Li, Lichi Li, Anh Dao, Xinyu Zhou, Wenjun Huang, Tianyi Ma, Yicheng Qiao, Zheda Mai 等 18 位作者
Visual large language models (VLLMs) are increasingly used as embodied agents, yet industrial warehouses demand more than reaching a goal: agents must navigate safely around moving workers and vehicles. To evaluate embodied agents' goal reaching and safety, we introduce IndustryNav, a simulation environment with 24 warehouse scenes for closed-loop navigation to specified destinations. Alongside task success and efficiency, we introduce safety-aware metrics that capture collisions and near-collision warnings along each trajectory. Experiments with eleven leading open- and closed-source VLLMs and three navigation baselines show that agents can reach their goals yet collide with moving obstacles, exposing a gap between success and safe navigation. This gap matters in safety-critical settings, where a single collision can end the task. IndustryNav also supports policy learning through automated oracle trajectories and simulator interaction, with held-out scenes for generalization tests. These experiments show that policies can succeed on new routes in familiar warehouses yet fail to transfer to held-out layouts, making cross-scene generalization an open challenge.
Yuanjie Lyu, Chengyu Wang, Haonan Zheng, Yuanhao Yue, Junbing Yan, Ming Wang, Jun Huang
Modern industrial applications increasingly demand language models that act as agents, capable of multi-step reasoning and tool use in real-world settings. These tasks are typically performed under strict cost and latency constraints, making small agentic models highly desirable. In this paper, we introduce the AgenticQwen family of models, trained via multi-round reinforcement learning (RL) on synthetic data and a limited amount of open-source data. Our training framework combines reasoning RL and agentic RL with dual data flywheels that automatically generate increasingly challenging tasks. The reasoning flywheel increases task difficulty by learning from errors, while the agentic flywheel expands linear workflows into multi-branch behavior trees that better reflect the decision complexity of real-world applications. We validate AgenticQwen on public benchmarks and in an industrial agent system. The models achieve strong performance on multiple agentic benchmarks, and in our industrial agent system, close the gap with much larger models on search and data analysis tasks.
尚无 AI 解读，展示原文摘要
cs.CL · cs.AI
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track) · 2026 · 正式发表
Wenrui Cai, Chengyu Wang Independent Researcher and Alibaba Group, Junbing Yan Alibaba Group, Jun Huang Alibaba Group and independent researcher, Xiangzhong Fang Shanghai Jiao Tong University
Tasks such as mathematical problem solving and coding require models to leverage chain-of-thought (CoT) processes, enabling human-like reasoning strategies. However, the advancement of large reasoning models (LRMs) is hindered by the lack of comprehensive CoT datasets. Existing resources often fail to provide extensive reasoning problems with coherent CoT processes distilled from multiple teacher models, and do not account for multifaceted properties describing the internal characteristics of CoTs.To address these challenges, we introduce OmniThought, a large-scale dataset featuring 2 million CoT processes generated and validated by multiple powerful LRMs. Each CoT process in OmniThought is annotated with novel Reasoning Verbosity (RV) and Cognitive Difficulty (CD) scores, which characterize the appropriateness of CoT verbosity and the cognitive difficulty level for models to comprehend these reasoning processes. We further establish a self-reliant pipeline to curate this dataset. Extensive experiments using Qwen2.5 and Qwen3 of various sizes demonstrate the positive impact of our RV and CD scores on LRM training effectiveness. Based on the OmniThought dataset, we train and release a series of high-performing LRMs with enhanced reasoning abilities and optimized CoT output length. Our contributions advance the development of LRMs across different scales for solving complex reasoning tasks.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Yuanjie Lyu Alibaba Group and University of Science and Technology of China, Chengyu Wang Independent Researcher and Alibaba Group, Lei Shen, Jun Huang Alibaba Group and independent researcher, Tong Xu University of Science and Technology of China
Small LLMs often struggle to match the agentic capabilities of large, costly models. While reinforcement learning can help, progress has been limited by two structural bottlenecks: existing open-source agentic training data are narrow in task variety and easily solved; real-world APIs lack diversity and are unstable for large-scale reinforcement learning rollout processes. We address these challenges with SYNTHAGENT, a framework that jointly synthesizes diverse tool-use training data and simulates complete environments. Specifically, a strong teacher model creates novel tasks and tool ecosystems, then rewrites them into intentionally underspecified instructions. This compels agents to actively query users for missing details. When handling synthetic tasks, an LLM-based user simulator provides user-private information, while a mock tool system delivers stable tool responses. For rewards, task-level rubrics are constructed based on required subgoals, user-agent interactions, and forbidden behaviors. Across 14 challenging datasets in math, search, and tool use, models trained on our synthetic data achieve substantial gains, with small models showing performance comparable to some larger baselines in certain domains.
尚无 AI 解读，展示原文摘要
cs.CL · cs.AI
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表