大模型推理能力

Learning to Break: Knowledge-Enhanced Reasoning in Multi-Agent Debate System

Haotian Wang, Xiyuan Du, Weijiang Yu, Qianglong Chen, Kun Zhu, Zheng Chu, Lian Yan, Yi Guan

专题命中 规划推理 ：reasoning(title)；分类 cs.CL、cs.AI

Comments 18 pages, 10 figures, work in progress

Planning with Logical Graph-based Language Model for Instruction Generation

Fan Zhang, Kebing Jin, Hankz Hankui Zhuo

专题命中 规划推理 ：planning(title)；分类 cs.CL、cs.AI

Comments 9 pages, 8 figures

In-Context Decision Transformer: Reinforcement Learning via Hierarchical Chain-of-Thought

Sili Huang, Jifeng Hu, Hechang Chen, Lichao Sun, Bo Yang

专题命中 规划推理 ：chain-of-thought(title)；分类 cs.AI、cs.LG

Transformer Based Planning in the Observation Space with Applications to Trick Taking Card Games

Douglas Rebstock, Christopher Solinas, Nathan R. Sturtevant, Michael Buro

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Improving Pacing in Long-Form Story Planning

Yichen Wang, Kevin Yang, Xiaoming Liu, Dan Klein

专题命中 规划推理 ：planning(title)；分类 cs.CL、cs.AI

Comments EMNLP Findings 2023

Few-Shot Table-to-Text Generation with Prompt Planning and Knowledge Memorization

Zhixin Guo, Minyxuan Yan, Jiexing Qi, Jianping Zhou, Ziwei He, Zhouhan Lin, Guanjie Zheng, Xinbing Wang

专题命中 规划推理 ：planning(title)；分类 cs.CL、cs.AI

Comments Accidental duplicate. Please see arXiv:2302.12468

Get Back Here: Robust Imitation by Return-to-Distribution Planning

Geoffrey Cideron, Baruch Tabanpour, Sebastian Curi, Sertan Girgin, Leonard Hussenot, Gabriel Dulac-Arnold, Matthieu Geist, Olivier Pietquin, Robert Dadashi

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

LookOut: Diverse Multi-Future Prediction and Planning for Self-Driving

Alexander Cui, Sergio Casas, Abbas Sadat, Renjie Liao, Raquel Urtasun

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

A Multi-Agent Reinforcement Learning Approach For Safe and Efficient Behavior Planning Of Connected Autonomous Vehicles

Songyang Han, Shanglin Zhou, Jiangwei Wang, Lynn Pepin, Caiwen Ding, Jie Fu, Fei Miao

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments This paper is submitted to IEEE Transactions on Intelligent Transportation Systems

Dynamic Planning in Open-Ended Dialogue using Reinforcement Learning

Deborah Cohen, Moonkyung Ryu, Yinlam Chow, Orgad Keller, Ido Greenberg, Avinatan Hassidim, Michael Fink, Yossi Matias, Idan Szpektor, Craig Boutilier, Gal Elidan

专题命中 规划推理 ：planning(title)；分类 cs.CL、cs.LG

Iterated Reasoning with Mutual Information in Cooperative and Byzantine Decentralized Teaming

Sachin Konan, Esmaeil Seraj, Matthew Gombolay

专题命中 规划推理 ：reasoning(title)；分类 cs.AI、cs.LG

Comments The first two authors contributed equally to this work (Published in ICLR 2022)

Journal ref International Conference on Learning Representations 2022

Safer Autonomous Driving in a Stochastic, Partially-Observable Environment by Hierarchical Contingency Planning

Ugo Lecerf, Christelle Yemdji-Tchassi, Pietro Michiardi

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments To appear in Generalizable Policy Learning in the Physical World workshop (ICLR 2022)

Meta-Reinforcement Learning for Heuristic Planning

Ricardo Luna Gutierrez, Matteo Leonetti

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments ICAPS 2021

Planning with Exploration: Addressing Dynamics Bottleneck in Model-based Reinforcement Learning

Xiyao Wang, Junge Zhang, Wenzhen Huang, Qiyue Yin

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments 15 pages, 8 figures

Discovering the Rationale of Decisions: Experiments on Aligning Learning and Reasoning

Cor Steging, Silja Renooij, Bart Verheij

专题命中 规划推理 ：reasoning(title)；分类 cs.AI、cs.LG

Comments 21 pages

Multi-Agent Safe Planning with Gaussian Processes

Zheqing Zhu, Erdem Bıyık, Dorsa Sadigh

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments 9 pages, 5 figures. Published at IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2020

Deceptive Reinforcement Learning for Privacy-Preserving Planning

Zhengshang Liu, Yue Yang, Tim Miller, Peta Masters

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

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

Deep reinforcement learning driven inspection and maintenance planning under incomplete information and constraints

C. P. Andriotis, K. G. Papakonstantinou

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Causality and Batch Reinforcement Learning: Complementary Approaches To Planning In Unknown Domains

James Bannon, Brad Windsor, Wenbo Song, Tao Li

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Path Planning Using Probability Tensor Flows

Francesco A. N. Palmieri, Krishna R. Pattipati, Giovanni Fioretti, Giovanni Di Gennaro, Amedeo Buonanno

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments Submitted for journal publication

Deep Visual Foresight for Planning Robot Motion

Chelsea Finn, Sergey Levine

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments ICRA 2017. Supplementary video: https://sites.google.com/site/robotforesight/

Bounded Planning in Passive POMDPs

Roy Fox, Naftali Tishby

专题命中 规划推理 ：planning(title)；分类 cs.AI、cs.LG

Comments Appears in Proceedings of the 29th International Conference on Machine Learning (ICML 2012)

A Reinforcement Learning-based Adaptive Control Model for Future Street Planning, An Algorithm and A Case Study

Qiming Ye, Yuxiang Feng, Jing Han, Marc Stettler, Panagiotis Angeloudis

专题命中 规划推理 ：planning(title,comments)；分类 cs.AI

Comments Proceeding for 57th ISOCARP World Planning Congress, Nov 8-11, 2021, Doha, Qatar

Mixed Integer Linear Programming For Exact Finite-Horizon Planning In Decentralized Pomdps

Raghav Aras, Alain Dutech, François Charpillet

专题命中 规划推理 ：planning(title,journal_ref)；分类 cs.AI

Journal ref Dans The International Conference on Automated Planning and Scheduling (2007)

Synergizing LLMs and Knowledge Graphs: A Novel Approach to Software Repository-Related Question Answering

融合大语言模型与知识图谱：一种针对软件仓库相关问答的新型方法

Samuel Abedu, SayedHassan Khatoonabadi, Emad Shihab

机构 * Data-driven Analysis of Software (DAS) Lab Department of Computer Science \& Software Engineering Concordia University Montreal QC Canada ； Concordia University 纠错

专题命中 规划推理 ：reasoning(abstract)；chain-of-thought(abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 本文提出融合大语言模型与知识图谱的方法，提升软件仓库问答的准确性与效率。

Comments Submitted to ACM Transactions on Software Engineering and Methodology for review

Cognitive-structured Multimodal Agent for Multimodal Understanding, Generation, and Editing

用于多模态理解、生成和编辑的认知结构多模态智能体

Feng Wang, Canmiao Fu, Zhipeng Huang, Chen Li, Jing Lyu, Ge Li

机构 * Peking University（北京大学） ； Tencent Inc（腾讯公司） 纠错

专题命中 规划推理 ：reasoning(abstract)；planning(abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 研究针对统一多模态模型在长时多模态对话中的局限，提出认知结构多模态智能体，通过外化视觉信息等方式改进。开发统一场景引擎，构建评估基准。实验显示该智能体检索准确率高、推理时间减半，还介绍了工具包，为长时多模态智能体提供新范式。

Comments 16 pages, 7 figures, 8 tables. Project page: https://caseclose.github.io/cma-harness/ Code: https://github.com/caseclose/cma-harness

BT-APE: A Computationally Light Backtracking Approach to Automatic Prompt Engineering for Requirements Classification

BT-APE: 一种用于需求分类的轻量级回溯自动提示工程方法

Mohammad Amin Zadenoori, Waad Alhoshan, Jacek Dąbrowski, Liping Zhao, Alessio Ferrari

专题命中 规划推理 ：CoT(abstract,abstract_cn)；chain-of-thought(abstract)

AI总结 提出轻量级回溯自动提示工程方法BT-APE，通过LLM生成候选、回溯搜索和动态示例选择迭代优化提示，在需求分类任务中以更低计算成本达到与资源密集型方法相当的准确率。

The Hidden Cost of Structured Generation in LLMs: Draft-Conditioned Constrained Decoding

结构生成在大语言模型中的隐性成本：草案条件约束解码

Avinash Reddy, Thayne T. Walker, James S. Ide, Amrit Singh Bedi

专题命中 规划推理 ：reasoning(abstract)；planning(abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 本文提出DCCD方法，通过分步解码提升结构生成的准确性与效率，实验显示在结构推理基准中，DCCD在严格结构准确性上提升了24个百分点，且在参数效率上优于传统约束解码。

Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense

下一代验证码：利用认知差距实现可扩展且多样化的GUI智能体防御

Jiacheng Liu, Yaxin Luo, Jiacheng Cui, Xinyi Shang, Xiaohan Zhao, Zhiqiang Shen

机构 * Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)（穆罕默德·本·扎耶德人工智能大学） ； MetaAgentX ； University College London（伦敦大学学院） 纠错

专题命中 规划推理 ：reasoning(abstract)；planning(abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 提出Next-Gen CAPTCHAs框架，通过利用人机在交互感知、记忆、决策和行动上的认知差距，设计动态任务以区分生物用户与AI智能体，实现可扩展的防御。

Comments Project page at https://greenoso.github.io/NextGen-CAPTCHAs_webpage/

S1-DeepResearch: Beyond Search, Toward Real-World Long-Horizon Research Agents

S1-DeepResearch：超越搜索，迈向真实世界的长周期研究智能体

Yao Dong, Xinglin Xiao, Liwei Dong, Xinlong Jin, Zhengbo Li, Heng Zhang, Duyun Wang, Nan Xu

机构 * XScience Lab（XScience实验室） ； Wenge AI（问格人工智能） 纠错

专题命中 规划推理 ：reasoning(abstract)；planning(abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 提出统一轨迹构建范式，结合封闭式问答与开放式探索，通过图基任务构建、智能体轨迹生成和多维验证，合成高质量长链推理轨迹，训练出在20个基准上达到开源最优的32B模型。