大模型推理能力

Neurosymbolic AI: The 3rd Wave

Artur d'Avila Garcez, Luis C. Lamb

专题命中 逻辑推理 ：reasoning(abstract)；logical reasoning(abstract)；分类 cs.AI、cs.LG

Comments 37 pages

Towards Analogy-Based Explanations in Machine Learning

Eyke Hüllermeier

专题命中 逻辑推理 ：reasoning(abstract)；logical reasoning(abstract)；分类 cs.AI、cs.LG

Comments 11 pages, 4 figures

History-Aware Question Answering in a Blocks World Dialogue System

Benjamin Kane, Georgiy Platonov, Lenhart K. Schubert

专题命中 逻辑推理 ：reasoning(abstract)；planning(abstract)；分类 cs.CL、cs.AI

Comments 16 pages, 4 figures

A Hybrid Knowledge-Grounded Framework for Safety and Traceability in Prescription Verification

一种混合知识引导的框架用于处方验证中的安全性和可追溯性

Yichi Zhu, Kan Ling, Xu Liu, Hengrun Zhang, Huiqun Yu, Guisheng Fan

机构 * School of Information Science and Engineering, East China University of Science and Technology（信息科学与工程学院，东华大学） 纠错

专题命中 逻辑推理 ：reasoning(abstract,comments)；logical reasoning(abstract)；分类 cs.AI

AI总结 本文提出PharmGraph-Auditor框架，通过混合制药知识库和知识引导的验证链，提升处方验证的安全性和可追溯性。

Comments 11 pages, 7 figures.Framework for safe prescription auditing and hybrid knowledge-grounded reasoning

Can LLMs Solve ASP Problems? Insights from a Benchmarking Study (Extended Version)

Lin Ren, Guohui Xiao, Guilin Qi, Yishuai Geng, Haohan Xue

机构 * School of Computer Science and Engineering, Southeast University, Nanjing, China（计算机科学与工程学院，东南大学，南京，中国） ； Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education, China（新一代人工智能技术及其跨学科应用关键实验室（东南大学），教育部，中国） 纠错

专题命中 逻辑推理 ：reasoning(abstract,comments)；logical reasoning(abstract)；分类 cs.AI

Comments Accepted for publication at the 22nd International Conference on Principles of Knowledge Representation and Reasoning (KR 2025). The code is available at https://github.com/HomuraT/ASPBench

Goal-Oriented Logic-based Semantic Communication for Neuro-Symbolic Reasoning with Applications onto Autonomous Driving

面向目标的基于逻辑的语义通信及用于自动驾驶的神经符号推理

Ahmet Faruk Saz, Duo Xu, Faramarz Fekri

专题命中 逻辑推理 ：reasoning(title)

AI总结 本文针对自动驾驶场景，提出基于一阶逻辑的目标导向语义通信方法，通过选择关键证据传输实现安全决策，在相同通信预算下优于均匀证据选择。

Chart Specification: Structural Representations for Incentivizing VLM Reasoning in Chart-to-Code Generation

图表规范：用于促进VLM在图表到代码生成中推理的结构表示

Minggui He, Mingchen Dai, Jian Zhang, Yilun Liu, Shimin Tao, Pufan Zeng, Osamu Yoshie, Yuya Ieiri

机构 * Waseda University（早稻田大学） ； University of Science and Technology of China（中国科学技术大学） ； NanKai University（南开大学） 纠错

专题命中 逻辑推理 ：reasoning(title)

AI总结 本文提出图表规范，通过结构化中间表示提升VLM在图表到代码生成中的结构保真度，实验显示其在数据效率和性能上优于现有方法。

Comments Accepted by Neurocomputing

EvoPlan: Evolutionary Neuro-Symbolic Robot Planning with Spatio-Temporal Guarantees

EvoPlan：具有时空保证的进化神经符号机器人规划

Bhavya Sai Nukapotula, Samin Moosavi, Haoze Wang, Luke Duncan, Diya Shakkottai, Varun Murali, Srinivas Shakkottai

机构 * Texas A&M University（德克萨斯农工大学） 纠错

专题命中 逻辑推理 ：planning(title)

AI总结 研究针对基于LLM的机器人规划器不足及经典PDDL规划器问题，提出含离线STL约束挖掘、进化PDDL规划器和受约束执行循环的神经符号框架，能保证计划属性，在基准测试中表现良好且可在无云环境下部署。

Neuro-Symbolic Reasoning for Vulnerability Detection

用于漏洞检测的神经符号推理

Yanjie Zhao, Hongjie Chen, Li Lu, Zhou Yang, Xiao Cheng, Haoyu Wang

专题命中 逻辑推理 ：reasoning(title)

AI总结 研究基于大语言模型的漏洞检测不可靠性，提出神经符号框架LeanGuard，通过将代码解释与安全义务判定分离，神经侧过滤事实，符号侧形式验证，证据感知裁决器权衡结果，在五个CWE类上实例化框架。

On the Timed Temporal Logic Planning of Coupled Multi-Agent Systems

关于耦合多智能体系统的时序时序逻辑规划

Alexandros Nikou, Dimitris Boskos, Jana Tumova, Dimos V. Dimarogonas

专题命中 逻辑推理 ：planning(title)

AI总结 本文提出一种自动控制器合成方法，用于满足耦合约束的多智能体系统。通过设计分布式抽象和形式验证技术，计算满足高阶任务的个体运行。

I Would If I Could: Reasoning about Dynamics of Actions in Multi-Agent Systems

若我能的话：多智能体系统中行动动态的推理

Rustam Galimullin, Hermine Grosinger, Munyque Mittelmann

专题命中 逻辑推理 ：reasoning(title)

AI总结 本文提出ATL-D和ATEL-D，用于建模智能体行动的动态授予与撤销过程及其对知识的影响，探讨了逻辑表达力、规范系统关系及计算复杂性。

Comments This is an extended version of the paper with the same title that will appear in KR 2026, and which contains a technical appendix with proof details

Enhancing Local Life Service Recommendation with Agentic Reasoning in Large Language Model

通过大语言模型中的代理推理增强本地生活服务推荐

Shiteng Cao, Xiaochong Lan, Yuwei Du, Jie Feng, Yinxing Liu, Xinlei Shi, Yong Li

专题命中 逻辑推理 ：reasoning(title)

AI总结 本文提出一种联合预测生活需求和服务推荐的框架，通过行为聚类和强化学习提升推荐准确性。

Human-AI Interaction: Evaluating LLM Reasoning on Digital Logic Circuit included Graph Problems, in terms of creativity in design and analysis

人类-人工智能交互：评估LLM在包含图问题的数字逻辑电路中的推理能力，就设计和分析的创造性而言

Yogeswar Reddy Thota, Setareh Rafatirad, Homayoun Houman, Tooraj Nikoubin

专题命中 逻辑推理 ：reasoning(title)

AI总结 评估LLM在数字逻辑电路问题中的推理能力，发现其在复杂问题上与官方答案存在差距，且易陷入教科书模板。

Meta-GPT: Decoding the Metasurface Genome with Generative Artificial Intelligence

Meta-GPT：用生成式人工智能解码元表面基因组

David Dang, Stuart Love, Meena Salib, Quynh Dang, Samuel Rothfarb, Mysk Alnatour, Andrew Salij, Hou-Tong Chen, Ho Wai, Lee, Wilton J. M. Kort-Kamp

专题命中 逻辑推理 ：chain-of-thought(abstract,comments)；分类 cs.CL、cs.AI、cs.LG；reasoning(comments)

AI总结 Meta-GPT通过METASTRINGS符号语言实现光子学设计，以生成式人工智能解码元表面基因组。

Comments Keywords: Physics-informed machine learning; Transformer models; Reinforcement learning; Chain-of-thought reasoning; Metasurfaces; Nanophotonics; Inverse design

DeepTx: Real-Time Transaction Risk Analysis via Multi-Modal Features and LLM Reasoning

Yixuan Liu, Xinlei Li, Yi Li

专题命中 逻辑推理 ：reasoning(title)

Comments Accepted to ASE'25

From Classical Rationality to Contextual Reasoning: Quantum Logic as a New Frontier for Human-Centric AI in Finance

Fabio Bagarello, Francesco Gargano, Polina Khrennikova

专题命中 逻辑推理 ：reasoning(title)

Comments 19 pages, 5 figures, preprint version. Forthcoming in: Journal of Quantum Economics and Finance

Comp-LTL: Temporal Logic Planning via Zero-Shot Policy Composition

Taylor Bergeron, Zachary Serlin, Kevin Leahy

专题命中 逻辑推理 ：planning(title)

Comments 16 pages, 11 figures. Updated to reflect additional results

Logic of Awareness in Agent's Reasoning

Yudai Kubono, Teeradaj Racharak, Satoshi Tojo

专题命中 逻辑推理 ：reasoning(title)

Comments A version reflecting the corrections of errors. The results are the same as originally stated. The erratum to arXiv:2309.09214v2 is attached at the end

Journal ref Proceedings of the 15th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART, 2023

Verse: A Python library for reasoning about multi-agent hybrid system scenarios

Yangge Li, Haoqing Zhu, Katherine Braught, Keyi Shen, Sayan Mitra

专题命中 逻辑推理 ：reasoning(title)

Comments 26 pages, 16 figures

Verification of Logical Consistency in Robotic Reasoning

Hongyang Qu, Sandor M. Veres

专题命中 逻辑推理 ：reasoning(title)

Journal ref Robotics and Autonomous Systems, Vol. 83(2016), 44-56

Modeling and Verification of Agent based Adaptive Traffic Signal using Symbolic Model Verifier

Vivek Vishal, Sagar Gugwad, Sanjay Singh

专题命中 逻辑推理 ：verifier(title)

Comments 13 pages, 6 figures, Submitted to International Journal of Computer Application (IJCA)

Reasoning about Games via a First-order Modal Model Checking Approach

Davi Romero de Vasconcelos, Edward Hermann Haeusler

专题命中 逻辑推理 ：reasoning(title)

Comments Extended version of article published in the SBMF 2007. Accepted to ENTCS. Withdrawn from ENTCS in 2014 in virtue to submission to other venue

Symbolic Planning and Control Using Game Theory and Grammatical Inference

Jie Fu, Herbert G. Tanner, Jeffrey Heinz, Jane Chandlee, Konstantinos Karydis, Cesar Koirala

专题命中 逻辑推理 ：planning(title)

Application of distributed constraint satisfaction problem to the agent-based planning in manufacturing systems

S. Kornienko, O. Kornienko, P. Levi

专题命中 逻辑推理 ：planning(title)

Journal ref Proceedings of the International Scientific Congress "Intelligent Systems (IEEE AIS'03)" and "Intelligent CAD's (CAD-2003)", p.124-140, Divnomorsk, Russia, 2003

Do Recipes Have Personas? Characterizing and Generating Creator Style in Attributed Procedural Graphs

食谱是否具有人设？在带属性的过程图中刻画与生成创作者风格

Lei Jiang

机构 * Microsoft（微软公司） 纠错

专题命中 逻辑推理 ：reasoning(abstract)；planning(abstract)；分类 cs.AI

AI总结 该研究针对大语言模型生成同质化逻辑的问题，构建了映射特定创作者的烹饪执行图数据集，提出结构化两阶段模型，可精准刻画与生成创作者的过程风格，集成方法能融合语义推理与拓扑控制优势。

Comments Accepted at the 29th International Conference on Discovery Science (DS 2026). 15 pages, 2 figures

LLM-Only PDDL Domain Repair with Open-Weight Models

仅使用大语言模型的PDDL领域修复：基于开放权重模型

Nader Karimi Bavandpour, Pascal Bercher

专题命中 逻辑推理 ：reasoning(abstract)；planning(abstract)；分类 cs.AI

AI总结 本文评估开放权重大语言模型仅用LLM方法修复PDDL模型错误的能力，发现其F1分数优于符号基线，但测试通过率仍不足，无法保证满足修复所需的测试约束。

SKILL: Self-correcting Knowledge-guided Iterative Large Language Model Agent for Logic Optimization

SKILL：用于逻辑优化的自校正知识引导迭代大语言模型智能体

Rui Yang

机构 * University of California, Riverside（加州大学河滨分校） 纠错

专题命中 逻辑推理 ：reasoning(abstract)；planning(abstract)；分类 cs.AI

AI总结 针对逻辑综合优化的挑战，提出SKILL智能体，结合多LLM推理与RL交互，经基准测试实现12.4%的PDA提升及86.3%的50万门级逻辑系统成功率。

Symbolic Attack Chain Generation from Atomic Red Team Techniques: An Empirical Study of Predicate Representation Granularity

基于Atomic Red Team技术的符号化攻击链生成：谓词表示粒度的实证研究

Ramya Varunsegar

专题命中 逻辑推理 ：reasoning(abstract)；planning(abstract)；分类 cs.AI

AI总结 本研究对比九类与五类谓词粒度的AALM，发现粒度对攻击链有效性影响小，81.3%结果一致，高粒度仅提升规划论证的内部结构分辨率。

From Passive Video to Editable Experience: Physically Grounded Experience Synthesis for Embodied Intelligence

从被动视频到可编辑体验：具身智能的物理基础体验合成

Jia Luo

专题命中 逻辑推理 ：planning(abstract)；verifier(abstract)；分类 cs.AI

AI总结 针对具身AI的数据瓶颈，提出Pegasus低资源框架，通过结构化知识传递将人类操作视频转化为机器人可学习数据，经多基准与机器人评估验证其跨具身翻译及数据生成的有效性。

AlphaOracle: Oracle bone script decipherment via human-workflow-inspired deep learning

AlphaOracle：通过受人工工作流程启发的深度学习进行甲骨文破译

Yuliang Liu, Haisu Guan, Pengjie Wang, Xinyu Wang, Jinpeng Wan, Kaile Zhang, Handong Zheng, Xingchen Liu, Zhebin Kuang, Huanxin Yang, Bang Li, Yongge Liu, Lianwen Jin, Xiang Bai

机构 * School of Software Engineering, Huazhong University of Science and Technology（华中科技大学软件学院） ； School of Electronic Information Engineering, South China University of Technology（华南理工大学电子信息学院） ； Key Laboratory of Oracle Bone Inscriptions Information Processing, Anyang Normal University（安阳师范学院甲骨文信息处理重点实验室） 纠错

专题命中 逻辑推理 ：reasoning(abstract)；logical reasoning(abstract)；分类 cs.CL

AI总结 针对约3000个未破译甲骨文字符的问题，提出受人工工作流程启发的AlphaOracle框架，经多阶段管道进行甲骨文破译，与专家解读高度一致，还能减少分析时间，为甲骨文及其他未破译文字研究提供参考。

Comments Accepted by The Innovation 2026

Journal ref The Innovation 7(11), 101462, 2026