大模型推理能力

Symbolic Logic meets Machine Learning: A Brief Survey in Infinite Domains

Vaishak Belle

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

Three Modern Roles for Logic in AI

Adnan Darwiche

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

Comments To be published in PODS 2020

Neuro-symbolic Architectures for Context Understanding

Alessandro Oltramari, Jonathan Francis, Cory Henson, Kaixin Ma, Ruwan Wickramarachchi

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

Comments In: Ilaria Tiddi, Freddy Lecue, Pascal Hitzler (eds.), Knowledge Graphs for eXplainable AI -- Foundations, Applications and Challenges. Studies on the Semantic Web, IOS Press, Amsterdam, 2020. arXiv admin note: text overlap with arXiv:1910.14087

Cost-Based Goal Recognition Meets Deep Learning

Mariane Maynard, Thibault Duhamel, Froduald Kabanza

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

Comments An earlier version of this paper was published in PAIR (AAAI 2019 workshop)

A Spoken Dialogue System for Spatial Question Answering in a Physical Blocks World

Georgiy Platonov, Benjamin Kane, Aaron Gindi, Lenhart K. Schubert

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

Comments 9 pages (with references), 2 figures

Engaging in Dialogue about an Agent's Norms and Behaviors

Daniel Kasenberg, Antonio Roque, Ravenna Thielstrom, Matthias Scheutz

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

Comments Accepted to the 1st Workshop on Interactive Natural Language Technology for Explainable Artificial Intelligence (NL4XAI)

Generating Justifications for Norm-Related Agent Decisions

Daniel Kasenberg, Antonio Roque, Ravenna Thielstrom, Meia Chita-Tegmark, Matthias Scheutz

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

Comments Accepted to the Proceedings of the 12th International Conference on Natural Language Generation (INLG 2019)

A Logic-Based Framework Leveraging Neural Networks for Studying the Evolution of Neurological Disorders

Francesco Calimeri, Francesco Cauteruccio, Luca Cinelli, Aldo Marzullo, Claudio Stamile, Giorgio Terracina, Francoise Durand-Dubief, Dominique Sappey-Marinier

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

Comments Under consideration in Theory and Practice of Logic Programming (TPLP)

Comparison of Artificial Intelligence Techniques for Project Conceptual Cost Prediction

Haytham H. Elmousalami

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

Comments arXiv admin note: text overlap with arXiv:1905.11804

LYRICS: a General Interface Layer to Integrate Logic Inference and Deep Learning

Giuseppe Marra, Francesco Giannini, Michelangelo Diligenti, Marco Gori

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

Comments To appear in proceedings of ECML PKDD 2019

Grounding Value Alignment with Ethical Principles

Tae Wan Kim, Thomas Donaldson, John Hooker

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

Human-aided Multi-Entity Bayesian Networks Learning from Relational Data

Cheol Young Park, Kathryn Blackmond Laskey

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

Incremental and Iterative Learning of Answer Set Programs from Mutually Distinct Examples

Arindam Mitra, Chitta Baral

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

Proceedings of the 2016 Workshop on Semantic Spaces at the Intersection of NLP, Physics and Cognitive Science

Dimitrios Kartsaklis, Martha Lewis, Laura Rimell

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

Journal ref EPTCS 221, 2016

Question Answering via Integer Programming over Semi-Structured Knowledge

Daniel Khashabi, Tushar Khot, Ashish Sabharwal, Peter Clark, Oren Etzioni, Dan Roth

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

Comments Extended version of the paper accepted to IJCAI'16

Towards AI-Complete Question Answering: A Set of Prerequisite Toy Tasks

Jason Weston, Antoine Bordes, Sumit Chopra, Alexander M. Rush, Bart van Merriënboer, Armand Joulin, Tomas Mikolov

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

Large Margin Nearest Neighbor Embedding for Knowledge Representation

Miao Fan, Qiang Zhou, Thomas Fang Zheng, Ralph Grishman

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

Comments arXiv admin note: text overlap with arXiv:1503.08155

Analysing the behaviour of robot teams through relational sequential pattern mining

Grazia Bombini, Raquel Ros, Stefano Ferilli, Ramon Lopez de Mantaras

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

Comments 25 pages

An Information-Flow Perspective on Explainability Requirements: Specification and Verification

可解释性需求的信息流视角：规范与验证

Bernd Finkbeiner, Hadar Frenkel, Julian Siber

机构 * CISPA Helmholtz Center for Information Security（信息安全研究中心） 纠错

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

AI总结 该研究从信息流视角，采用扩展反事实因量化的认知时态逻辑，提出可解释性需求的规范与验证方法，实现有限状态模型的可解释性检查，可区分可解释与不可解释系统并支持隐私需求设定。

Comments This is an extended and corrected version of the paper presented at the 22nd International Conference on Principles of Knowledge Representation and Reasoning (KR 2025); see the appendix for details

Same Formulas, Different Semantics: Do Language Models Follow Modal Logic Specifications?

相同公式，不同语义：语言模型是否遵循模态逻辑规范？

Réemi Andrieu, Damien Sileo

机构 * Univ. Lille（里尔大学） ； Inria（法国国家信息与自动化研究所） ； CNRS（法国国家科学研究中心） ； Centrale Lille（里尔中央理工学院） ； UMR 9189 - CRIStAL（UMR 9189 - CRIStAL（法国国家科研中心联合研究机构）） 纠错

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

AI总结 该研究探究语言模型是否遵循模态逻辑规范，构建成对模态问题测试五款模型，发现遵循规定模态语义依赖推理模式与模型身份，发布相关产物。

Comments 9 pages. Code: https://github.com/sileod/modal-semantics-reasoning. Data and artifacts: https://huggingface.co/datasets/sileod/modal-semantics-reasoning

Teaching LLMs String Matching, Backtracking, and Error Recovery to Deduce Bases and Truth Tables for the Combinatorially Exploding Bit Manipulation Puzzles

教LLM字符串匹配、回溯与错误恢复：为组合爆炸的位操作谜题推导基和真值表

Prateek Agnihotri, Sanchit Jain, Prabhat Agnihotri, Aditya Prasad, Shubham Jain

机构 * NVIDIA 纠错

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

AI总结 提出一种新方法，通过字符串相似性、结构化搜索和自主错误恢复，避免复杂布尔逻辑，解决位操作谜题中的组合爆炸问题，实现96%以上验证准确率。

Comments 22 pages, 4 figures, 2 tables. 7th Place Solution for the NVIDIA Nemotron Model Reasoning Challenge (Kaggle)

A Scalable Approach to Probabilistic Neuro-Symbolic Robustness Verification

Vasileios Manginas, Nikolaos Manginas, Edward Stevinson, Sherwin Varghese, Nikos Katzouris, Georgios Paliouras, Alessio Lomuscio

机构 * Department of Computer Science and Leuven.AI KU Leuven Belgium（比利时列日大学计算机科学系） ； Department of Computing Imperial College London UK（伦敦帝国学院计算机系） 纠错

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

Comments 19th Conference on Neurosymbolic Learning and Reasoning

Counterfactuals Modulo Temporal Logics

Bernd Finkbeiner, Julian Siber

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

Comments 24th International Conference on Logic for Programming, Artificial Intelligence and Reasoning (LPAR-23)

Using Explainable Scheduling for the Mars 2020 Rover Mission

Jagriti Agrawal, Amruta Yelamanchili, Steve Chien

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

Comments Submitted to the International Workshop of Explainable AI Planning (XAIP) at the International Conference on Automated Planning and Scheduling (ICAPS) 2020

Analyzing Differentiable Fuzzy Implications

Emile van Krieken, Erman Acar, Frank van Harmelen

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

Comments 10 pages, 10 figures, accepted to 17th International Conference on Principles of Knowledge Representation and Reasoning (KR 2020). arXiv admin note: substantial text overlap with arXiv:2002.06100

A Framework for Combining Defeasible Argumentation with Labeled Deduction

Carlos Iván Chesñevar, Guillermo Ricardo Simari

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

Comments 15 pages, presented at CMSRA Workshop 2003. Buenos Aires, Argentina

Journal ref In "Computer Modeling of Scientific Reasoning" (C.Delrieux, J.Legris, Eds.). Pp. 43-56, Ed. Ediuns, Argentina, 2003. ISBN 987-89281-89-6

VISA: Agentic Self-Evolving Data Synthesis for Multimodal Instruction Following

VISA：用于多模态指令遵循的智能体式自进化数据合成

Min Zeng, Guanxin Tan, Libin Cen, Yafei Wen, Rui Hu, Liuyang Bian, Xiaolong Chen, Xiaoxin Chen

机构 * vivo AI Lab（vivo AI实验室） 纠错

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

AI总结 VISA是一种智能体式自进化数据合成框架，通过自进化循环优化多模态指令合成，在MM-IFEval等基准上提升了多模态指令遵循性能。

Why Knowing Both Hops Is Not Enough: Understanding Two-Hop Generalization in Language Models

仅知道两跳信息是不够：理解语言模型中的两跳泛化

Zili Zhang, Yilin Wang, Heng Wang, Herun Wan, Minnan Luo

机构 * Xi’an Jiaotong University（西安交通大学） ； University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校） 纠错

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

AI总结 本研究针对语言模型两跳查询失败问题，在受控环境训练Transformer，揭示其泛化规律与跨层不匹配机制，提出循环式训练策略以提升分布外两跳泛化能力。

Comments EMNLP 2026, findings

Large Language Model Few-Shot Prompting with Dilemma Training Outperforms Human Surrogates in Predicting Patient Preferences

采用困境训练的大语言模型少样本提示在预测患者偏好方面优于人类替代者

Natasha Ureyang, Sebastian Porsdam Mann, Yuxin Liu, Zuriel Hassirim, Melanie Almonte, Wenhao Chen, Joyce Ng, Thant Nay Lin, Aung Thiha, Gerald CH Koh, Brian David Earp, Pin Sym Foong

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

AI总结 该研究提出采用困境训练的P4-DT智能体，通过12组患者-替代者配对实验，其预测患者治疗选择的准确率达81.7%，优于人类替代者及相关方法。

Vibe Compiler: A Research-Logic Synthesis Tool That Runs without Prompt Engineering -Toward Enhancing Metacognition for Sustaining Agency in the Age of Generative AI-

Vibe Compiler：无需提示工程即可运行的研究逻辑综合工具——迈向生成式AI时代维持智能体的元认知

Riichiro Mizoguchi, Tomoki Aburatani, Kento Koike, Machi Shimmei

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

AI总结 针对生成式AI可能侵蚀人类认知智能体的问题，提出基于综合-分析互惠模型的Vibe Compiler工具，通过16个学术参数的本体编译研究想法，以反思性问题引导研究人员，减少对复杂提示的依赖，提升AI辅助推理效果。

Comments 98 pages, 3 figures. English version (pp. 1-54) followed by a Japanese version (pp. 55-98)