AI Agent

Video Language Planning

Yilun Du, Mengjiao Yang, Pete Florence, Fei Xia, Ayzaan Wahid, Brian Ichter, Pierre Sermanet, Tianhe Yu, Pieter Abbeel, Joshua B. Tenenbaum, Leslie Kaelbling, Andy Zeng, Jonathan Tompson

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments https://video-language-planning.github.io/

Compositional Foundation Models for Hierarchical Planning

Anurag Ajay, Seungwook Han, Yilun Du, Shuang Li, Abhi Gupta, Tommi Jaakkola, Josh Tenenbaum, Leslie Kaelbling, Akash Srivastava, Pulkit Agrawal

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments Website: https://hierarchical-planning-foundation-model.github.io/

Integrating Symmetry into Differentiable Planning with Steerable Convolutions

Linfeng Zhao, Xupeng Zhu, Lingzhi Kong, Robin Walters, Lawson L. S. Wong

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments ICLR 2023 camera-ready version. Original name = "Integrating Symmetry into Differentiable Planning". Website: http://lfzhao.com/SymPlan

Learning to Operate in Open Worlds by Adapting Planning Models

Wiktor Piotrowski, Roni Stern, Yoni Sher, Jacob Le, Matthew Klenk, Johan deKleer, Shiwali Mohan

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG；autonomous agent(comments)

Comments To appears in the Proceedings of the 22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2023)

Planning with Diffusion for Flexible Behavior Synthesis

Michael Janner, Yilun Du, Joshua B. Tenenbaum, Sergey Levine

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments ICML 2022 (long talk). Project page and code at https://diffusion-planning.github.io/

Is Policy Learning Overrated?: Width-Based Planning and Active Learning for Atari

Benjamin Ayton, Masataro Asai

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments Accepted in ICAPS 2022, Planning & Learning track. Edits: Title change. Fixed the variance update formula. Updated the scores of $π$-IW as per authors' request. Included the results of DQN and EfficientZero

Deceptive Reinforcement Learning for Privacy-Preserving Planning

Zhengshang Liu, Yue Yang, Tim Miller, Peta Masters

专题命中 规划决策 ：planning(title)；agent(abstract)；分类 cs.AI、cs.LG；autonomous agent(journal_ref)

Journal ref Proceedings of the 20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2021)

Offline Learning for Planning: A Summary

Giorgio Angelotti, Nicolas Drougard, Caroline Ponzoni Carvalho Chanel

专题命中 规划决策 ：planning(title,comments)；autonomous agent(abstract)；分类 cs.AI、cs.LG

Comments 9 pages, ICAPS 2020 Conference - Bridging the Gap Between AI Planning and Reinforcement Learning (PRL) Workshop

Integrating Acting, Planning and Learning in Hierarchical Operational Models

Sunandita Patra, James Mason, Amit Kumar, Malik Ghallab, Paolo Traverso, Dana Nau

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments Accepted in ICAPS 2020 (30th International Conference on Automated Planning and Scheduling)

Hierarchical Reinforcement Learning with Abductive Planning

Kazeto Yamamoto, Takashi Onishi, Yoshimasa Tsuruoka

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments 7 pages, 6 figures, ICML/IJCAI/AAMAS 2018 Workshop on Planning and Learning (PAL-18)

Information-Theoretic Approach to Efficient Adaptive Path Planning for Mobile Robotic Environmental Sensing

Kian Hsiang Low, John M. Dolan, Pradeep Khosla

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments 19th International Conference on Automated Planning and Scheduling (ICAPS 2009), Extended version with proofs, 11 pages

Multi-Robot Informative Path Planning for Active Sensing of Environmental Phenomena: A Tale of Two Algorithms

Nannan Cao, Kian Hsiang Low, John M. Dolan

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG；autonomous agent(comments)

Comments 12th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013), Extended version with proofs, 15 pages

Active Markov Information-Theoretic Path Planning for Robotic Environmental Sensing

Kian Hsiang Low, John M. Dolan, Pradeep Khosla

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG；autonomous agent(comments)

Comments 10th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2011), Extended version with proofs, 11 pages

Gaussian Process Bandits for Tree Search: Theory and Application to Planning in Discounted MDPs

Louis Dorard, John Shawe-Taylor

专题命中 规划决策 ：planning(title,abstract)；分类 cs.AI、cs.LG

Comments Second draft. Tried to follow the JMLR formatting guidelines. Made corrections to the section on planning in MDPs

AEGIS: Preventing Cross-Domain Resource Abuse in MCP

AEGIS：防止MCP中的跨域资源滥用

Shriti Priya, Teryl Taylor, Frederico Araujo

专题命中 规划决策 ：agent(summary_cn,abstract)；分类 cs.AI

AI总结 本文提出AEGIS，该组件借助LLM将MCP工具调用归一化为统一表示，结合Open Policy Agent与ContextForge AI Gateway，可防止跨异构MCP工具和模态的资源滥用。

SHE: Trajectory-driven Safety Harness Evolution for LLM Agents

SHE：面向大语言模型智能体的轨迹驱动安全管控机制演化

Wanying Qu, Qinghua Mao, Yu Li, Jiyao Liu, Xin Zhang, Dadi Guo, Yanxu Zhu, Qingyu Liu, Leitao Yuan, Xi Lin, Shanfeng Zhu, Yanwei Fu, Jing Shao, Xia Hu, Dongrui Liu

机构 * Shanghai Artificial Intelligence Laboratory（上海人工智能实验室） ； Fudan University（复旦大学） ； Shanghai Jiao Tong University（上海交通大学） ； The Hong Kong University of Science and Technology（香港科技大学） 纠错

专题命中 规划决策 ：agent(summary_cn,abstract)；分类 cs.AI

AI总结 本研究提出SHE框架，将LLM智能体的安全管控机制分解为四个构件并引入归因引导演化循环，在Agent-SafetyBench上使攻击成功率降低3.1倍，还具备泛化与跨模型迁移能力。

Comments Project: https://github.com/RainbowQTT/SHE

A game theory for foundation models shows new paths to rational cooperation through similarity inference

基础模型的博弈论：通过相似性推理实现理性合作的新路径

Alexander Meulemans, Maciej Wołczyk, Marissa A. Weis, Rajai Nasser, Roberta Rocca, Seijin Kobayashi, Guillaume Lajoie, Angelika Steger, Blake Richards, Marcus Hutter, James Manyika, Rif A. Saurous, João Sacramento, Blaise Agüera y Arcas

专题命中 规划决策 ：agent(abstract)；AI agent(abstract)；autonomous agent(abstract)；planning(abstract)

AI总结 本文提出针对基础模型智能体的嵌入贝叶斯智能体理论模型，通过嵌入均衡机制实现相似性推理，破解经典博弈论预测的社会困境中相互背叛问题，达成稳定合作。

Comments 75 pages, 11 figures

Partner Capability Estimation for Task-Agnostic Adaptation in Ad-Hoc Teamwork

面向临时团队中任务无关适应的伙伴能力估计

Peter Tisnikar, Maja Swieczkowska, Benteng Ma, Gerard Canal, Matteo Leonetti

机构 * King’s College London（伦敦国王学院） 纠错

专题命中 规划决策 ：agent(abstract)；autonomous agent(abstract)；planning(abstract)；multi-agent(abstract)

AI总结 该研究针对临时团队协作中伙伴能力隐藏与人类行为不可预测问题，提出CE-CM及CE-CM-Div方法，通过仿真采样与多样规划器rollout实现任务无关的能力估计，提升了团队协作的可行性与鲁棒性。

Comments 44 pages, 18 figures, submitted

Context Assembly as the Controlled Variable: A Control-Theoretic View of Harness Policies for Frozen LLM Agents

作为受控变量的上下文组装：冻结大语言模型智能体利用策略的控制理论视角

Debjyoti Paul

专题命中 规划决策 ：agent(abstract)；planning(abstract)；agentic(abstract)；multi-agent(abstract)

AI总结 研究聚焦大语言模型智能体，以往控制工具选择等，本文将上下文组装视为受控变量，通过上下文博弈或强化学习策略在线学习，进行形式分解、稳定性论证及不确定性校准分析，给出控制理论视角的相关证据。

Comments 6 pages, 2 figures, 1 table. Code and companion paper's data: https://github.com/dpaul0501/context-optimization-rl

PlanCraft: Sketch, Refine, and Furnish for Architect-Inspired Progressive 3D Residential Scene Generation

PlanCraft：用于受建筑师启发的渐进式3D住宅场景生成的草图绘制、细化和布置

Pengyu Zeng, Yuqin Dai, Jun Yin, Ziyang Han, Ng Cheuk Hei, Jing Zhong, Chaoyang Shi, ZhanXiang Jin, Maowei Jiang, Yuxing Han, Shuai Lu

专题命中 规划决策 ：agent(summary_cn,abstract)；分类 cs.CL

AI总结 研究针对自动住宅平面图生成中忽视的设计渐进性及二维平面图重要性问题，提出PlanCraft，通过SketchPlan提供训练信号，PlanCraft-Diff细化草图，PlanCraft-Agent布置场景，实验显示其在FID及空间合理性上优于现有方法。

NVAITC AI Scientist: A Governed End-to-End Research System -- A Hypertension GWAS Case Study

NVAITC人工智能科学家：一个受治理的端到端研究系统——以高血压全基因组关联研究为例

Eddie Huang, Ken Liao, Iven Fu, Yang-Hsien Lin, Chao-Shun Zhan, Andy Liao, Virginia Chen, Johnson Sun, Pika Wang, Richard Huang, Jiun-Cheng Jiang, Ting-Yuan Liu, Hsing-Fang Lu, Ray Y. Lee, Chi-Chou Liao, Simon See, Fuu-Jen Tsai

机构 * Department of Medical Research, China Medical University Hospital（中国医药大学附设医院医学研究部） ； Master Program for Digital Health Innovation, China Medical University（中国医药大学数字健康创新硕士项目） ； Laboratory for Statistical and Translational Genetics, RIKEN Center for Integrative Medical Sciences（理化学研究所综合医学科学中心统计与转化遗传学实验室） ； AI-Driven Genomic Medicine and Drug Discovery Lab, China Medical University Hospital（中国医药大学附设医院人工智能驱动的基因组医学与药物发现实验室） 纠错

专题命中 规划决策 ：agent(abstract)；planning(abstract)；workflow(abstract)；agentic(abstract)

AI总结 研究以高血压GWAS为例，介绍受治理的端到端智能研究系统NAIS，其集成多种功能。通过真实数据验证，NAIS能支持生物医学发现，规划队列提取等，重现高血压基因座，在药物性肝损伤预测中也有成果，可产生与专家主导流程相当的输出。

Comments 22 pages, 6 figures, 4 tables

Interactive Learning for LLM Reasoning

为LLM推理设计的交互学习

Hehai Lin, Shilei Cao, Sudong Wang, Haotian Wu, Minzhi Li, Linyi Yang, Juepeng Zheng, Chengwei Qin

机构 * The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州）） ； Sun Yat-sen University（中山大学） ； Southern University of Science and Technology（南方科技大学） ； National University of Singapore（新加坡国立大学） 纠错

专题命中 规划决策 ：agent(summary_cn,abstract)；multi-agent(abstract)；分类 cs.AI

AI总结 本文提出ILR框架，通过动态交互和感知校准提升LLM独立问题解决能力，实验表明其在多个基准测试中优于单agent学习。

Comments The code is available at https://github.com/linhh29/Interactive-Learning-for-LLM-Reasoning

OdysseyArena: Benchmarking Large Language Models For Long-Horizon, Active and Inductive Interactions

OdysseyArena: 为长视界、主动和归纳交互评估大型语言模型

Hang Yan, Fangzhi Xu, Qiushi Sun, Jinyang Wu, Zixian Huang, Muye Huang, Jingyang Gong, Zichen Ding, Kanzhi Cheng, Yian Wang, Xinyu Che, Zeyi Sun, Jian Zhang, Zhangyue Yin, Haoran Luo, Ben Kao, Qika Lin

机构 * National University of Singapore（新加坡国立大学） 纠错

专题命中 规划决策 ：agent(abstract)；autonomous agent(abstract)；planning(abstract)；agentic(abstract)

AI总结 本文提出OdysseyArena，通过长视界、主动和归纳交互评估大型语言模型，提供120个任务测量归纳效率和长视界发现，并通过OdysseyArena-Challenge测试极端交互视界下的模型稳定性，揭示前沿模型在复杂环境中的归纳能力瓶颈。

Comments 34 pages

Can LLM Agents Sustain Long-Horizon Organizational Dynamics?

LLM智能体能否维持长期组织动态？

Xuancheng Zhu, Yang Yue, Shuaibing Wan, Zihan Dou, Xiaohan Zhang, Yongrui Liu, Guoshun Nan

机构 * Beijing University of Posts and Telecommunications（北京邮电大学） 纠错

专题命中 规划决策 ：agent(abstract)；planning(abstract)；agentic(abstract)；multi-agent(abstract)

AI总结 提出TaskWeave分层智能体框架，通过记忆中心的协调机制（规划-分解-诊断-对齐循环和依赖感知追踪记忆）实现长期组织模拟，实验表明该框架能维持连贯的组织动态并产生可靠的人工制品。

Security of OpenClaw Agents: Fundamentals, Attacks, and Countermeasures

OpenClaw 代理的安全性：基础、攻击与对策

Yuntao Wang, Jianle Ba, Han Liu, Yanghe Pan, Jintao Wei, Zhou Su, Tom H. Luan, Linkang Du

机构 * School of Cyber Science and Engineering, Xi'an Jiaotong University（西安交通大学网络科学与工程学院） 纠错

专题命中 规划决策 ：agent(abstract)；AI agent(abstract)；autonomous agent(abstract)；multi-agent(abstract)

AI总结 本文综述了 OpenClaw 代理的安全挑战，分类分析了技能投毒、认知操纵、多代理级联故障和供应链漏洞等威胁，并总结了现有防御机制。

Comments 17 pages, 13 figures

Qumus: Realization of An Embodied AI Quantum Material Experimentalist

Qumus: 一种具身人工智能量子材料实验家的实现

Lihan Shi, Zhaoyi Joy Zheng, Xinzhe Juan, Yimin Wang, Ming Yin, Mayank Sengupta, Kristina Wolinski, Yanyu Jia, Jingzhi Shi, Derek Saucedo, Neill Saggi, Haosen Guan, Kenji Watanabe, Takashi Taniguchi, Ali Yazdani, Mengdi Wang, Sanfeng Wu

专题命中 规划决策 ：agent(abstract)；planning(abstract)；agentic(abstract)；multi-agent(abstract)

AI总结 本文提出Qumus，首个能够进行真实世界科学发现的具身人工智能量子材料实验家，通过机器人微型实验室实现了原子薄二维材料和范德瓦耳斯结构的制备与纳米加工，首次实现了AI生成石墨烯和原子薄场效应晶体管的AI制造。

Comments 29 Pages in total. Supplementary Demo Videos are available at https://qumus.ai

EigentSearch-Q+: Enhancing Deep Research Agents with Structured Reasoning Tools

EigentSearch-Q+: 通过结构化推理工具增强深度研究代理

Boer Zhang, Mingyan Wu, Dongzhuoran Zhou, Yuqicheng Zhu, Wendong Fan, Puzhen Zhang, Zifeng Ding, Guohao Li, Yuan He

机构 * Meta ； Northeastern University, China（东北大学） ； University of Oslo（奥斯陆大学） ； Bosch Center for AI（博世人工智能中心） ； University of Stuttgart（斯图加特大学） ； University of Cambridge（剑桥大学） ； Mina AI ； Amazon（亚马逊） 纠错

专题命中 规划决策 ：agent(abstract)；AI agent(abstract)；planning(abstract)；multi-agent(abstract)

AI总结 本文提出Q+工具，通过引导查询规划和证据提取提升深度研究代理的搜索效率，实验显示在多个基准测试中提升了模型准确性。

Documentless Assessments Using Nominal Group Interviews

无需文档的评估使用名义小组访谈

Eduardo Miranda

专题命中 规划决策 ：planning(summary_cn,abstract)；分类 cs.SE

AI总结 本文提出一种小组访谈技术，用于支持无文档的过程评估，同时促进参与者之间的协作。该方法借鉴敏捷软件开发中的用户故事概念，将CMMI具体实践转化为具体术语，并采用Planning Poker技术替代文档审查和审计访谈进行事实发现和验证。

Journal ref Software Quality Professional, March 2017, Volume 19 Issue 2

SCALAR: Learning and Composing Skills through LLM Guided Symbolic Planning and Deep RL Grounding

SCALAR：通过LLM引导的符号规划和深度RL接地学习与组合技能

Renos Zabounidis, Yue Wu, Simon Stepputtis, Woojun Kim, Yuanzhi Li, Tom Mitchell, Katia Sycara

专题命中 规划决策 ：planning(title,abstract)；分类 cs.LG；agent(comments)

AI总结 SCALAR通过结合LLM引导的符号规划与深度RL，实现技能学习与组合，显著提升任务完成效率和鲁棒性。

Comments Best Paper Award Honorable Mention at NeurIPS 2025 Workshop on Bridging Language, Agent, and World Models for Reasoning and Planning

Yunque DeepResearch Technical Report

Yunque DeepResearch 技术报告

Yuxuan Cai, Xinyi Lai, Peng Yuan, Weiting Liu, Huajian Li, Mingda Li, Xinghua Wang, Shengxie Zheng, Yanchao Hao, Yuyang Yin, Zheng Wei

机构 * Tencent BAC（腾讯BAC） ； Tsinghua University（清华大学） ； Fudan University（复旦大学） 纠错

专题命中 规划决策 ：agent(abstract)；autonomous agent(abstract)；agentic(abstract)；multi-agent(abstract)

AI总结 Yunque DeepResearch提出了一种分层、模块化且稳健的框架，通过多代理协调、动态上下文管理和主动监督模块解决深度研究中的关键挑战，实现先进性能并开源支持社区发展。