arXivDaily arXiv每日学术速递 周一至周五更新

AI 大模型

大模型推理能力

大模型数学、逻辑、规划、多步推理和测试时计算能力。

共收录 3046 信号源:cs.CL, cs.AI, cs.LG

1. 逻辑推理 3046 篇

2302.07257 2023-02-15 cs.CV eess.IV 67%

ChatCAD: Interactive Computer-Aided Diagnosis on Medical Image using Large Language Models

Sheng Wang, Zihao Zhao, Xi Ouyang, Qian Wang, Dinggang Shen

专题命中 逻辑推理 :reasoning(abstract);logical reasoning(abstract)

详情

展开后加载摘要…

URL PDF HTML 收藏
2202.12205 2022-12-16 cs.AI cs.CL cs.LG 67%

Is Neuro-Symbolic AI Meeting its Promise in Natural Language Processing? A Structured Review

Kyle Hamilton, Aparna Nayak, Bojan Božić, Luca Longo

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

Comments Survey

Journal ref Semantic Web, vol. Pre-press, no. Pre-press, pp. 1-42, 2022

详情

展开后加载摘要…

URL PDF HTML 收藏
2202.13758 2022-12-13 cs.CL cs.AI cs.CY cs.LG cs.LO 67%

Logical Fallacy Detection

Zhijing Jin, Abhinav Lalwani, Tejas Vaidhya, Xiaoyu Shen, Yiwen Ding, Zhiheng Lyu, Mrinmaya Sachan, Rada Mihalcea, Bernhard Schölkopf

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

Comments EMNLP 2021 Findings

详情

展开后加载摘要…

URL PDF HTML 收藏
2107.12079 2021-10-18 cs.CL cs.AI cs.IR cs.LG 67%

An Argumentative Dialogue System for COVID-19 Vaccine Information

Bettina Fazzinga, Andrea Galassi, Paolo Torroni

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

Comments 9 pages, 2 figures, Accepted at CLAR 2021

Journal ref Logic and Argumentation (2021). Lecture Notes in Computer Science, vol 13040

详情

展开后加载摘要…

URL PDF HTML 收藏
2105.09973 2021-10-01 cs.RO 67%

Mediating between Contact Feasibility and Robustness of Trajectory Optimization through Chance Complementarity Constraints

Luke Drnach, John Z. Zhang, Ye Zhao

专题命中 逻辑推理 :reasoning(abstract);planning(abstract)

Comments submitted to Frontiers in Robotics and AI

详情

展开后加载摘要…

URL PDF HTML 收藏
2003.00330 2021-06-15 cs.AI cs.CL cs.LG cs.LO 67%

Graph Neural Networks Meet Neural-Symbolic Computing: A Survey and Perspective

Luis C. Lamb, Artur Garcez, Marco Gori, Marcelo Prates, Pedro Avelar, Moshe Vardi

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

Comments Updated version, draft of accepted IJCAI2020 Survey Paper

详情

展开后加载摘要…

URL PDF HTML 收藏
1912.13283 2020-11-20 cs.CL cs.AI cs.LG 67%

oLMpics -- On what Language Model Pre-training Captures

Alon Talmor, Yanai Elazar, Yoav Goldberg, Jonathan Berant

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

Comments TACL 2020

详情

展开后加载摘要…

URL PDF HTML 收藏
2006.06609 2020-11-17 cs.CL cs.AI cs.LG 67%

Leap-Of-Thought: Teaching Pre-Trained Models to Systematically Reason Over Implicit Knowledge

Alon Talmor, Oyvind Tafjord, Peter Clark, Yoav Goldberg, Jonathan Berant

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

Comments Presented as Spotlight at NeurIPS 2020

详情

展开后加载摘要…

URL PDF HTML 收藏
1907.03950 2019-11-26 cs.AI cs.CL cs.CV cs.LG 67%

Learning by Abstraction: The Neural State Machine

Drew A. Hudson, Christopher D. Manning

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

Comments Published as a conference paper at NeurIPS 2019 (spotlight)

详情

展开后加载摘要…

URL PDF HTML 收藏
1911.02362 2019-11-11 q-bio.NC 67%

Bayesian Analogical Cybernetics

Adam Safron

专题命中 逻辑推理 :reasoning(abstract);logical reasoning(abstract)

详情

展开后加载摘要…

URL PDF HTML 收藏
1610.03935 2016-10-14 cs.LO 67%

Optimizing Epistemic Model Checking using Conditional Independence

Ron van der Meyden

专题命中 逻辑推理 :reasoning(abstract);logical reasoning(abstract)

详情

展开后加载摘要…

URL PDF HTML 收藏
1104.3216 2011-04-19 cs.DB 67%

Tuffy: Scaling up Statistical Inference in Markov Logic Networks using an RDBMS

Feng Niu, Christopher Ré, AnHai Doan, Jude Shavlik

专题命中 逻辑推理 :reasoning(abstract);logical reasoning(abstract)

Comments VLDB2011

Journal ref Proceedings of the VLDB Endowment (PVLDB), Vol. 4, No. 6, pp. 373-384 (2011)

详情

展开后加载摘要…

URL PDF HTML 收藏
0906.3769 2009-12-01 cs.CR 67%

Using Agent to Coordinate Web Services

C. H. Liu, Y. F. Lin, Jason J. Y. Chen

专题命中 逻辑推理 :reasoning(abstract);logical reasoning(abstract)

Comments 7 pages, International Journal of Computer Science and Information Security

Journal ref IJCSIS, June 2009, Vol. 2

详情

展开后加载摘要…

URL PDF HTML 收藏
2607.13115 2026-07-16 cs.AI cs.LG 新提交 66%

Improving Molecular Property Prediction in Small Language Models Using Graph-based Tools

使用基于图的工具改进小语言模型中的分子性质预测

Konstantinos Bougiatiotis, Dimitrios Kelesis, Georgios Paliouras

机构 * Institute of Informatics and Telecommunications, National Center for Scientific Research “Demokritos(信息与电信研究所,国家科学研究中心“德谟克利特”)

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

AI总结 研究针对小语言模型在分子性质预测时的结构盲目性问题,提出模块化上下文增强提示框架,通过GNN专家模型和提取子图来辅助预测,在MUTAG和Tox21数据集上实验表明该方法能显著提升预测准确性,验证了基序相关性,但与专门GNN模型仍有差距。

Comments Presented at the 2nd Causal Neuro-symbolic Artificial Intelligence (Causal NeSy): Toward Agentic LLMs with Neuro-Symbolic and Graph Based Reasoning Workshop @ ESWC2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.12981 2025-07-15 cs.AI cs.CL cs.IR 66%

SymRAG: Efficient Neuro-Symbolic Retrieval Through Adaptive Query Routing

Safayat Bin Hakim, Muhammad Adil, Alvaro Velasquez, Houbing Herbert Song

机构 * Department of Information Systems, University of Maryland, Baltimore County, USA(信息系统系,马里兰大学巴尔的摩县分校) Department of Computer Science and Engineering, University at Buffalo, USA(计算机科学与工程系,布法罗大学) Department of Computer Science, University of Colorado Boulder, USA(计算机科学系,科罗拉多大学波德分校)

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

Comments Accepted at 19th International Conference on Neurosymbolic Learning and Reasoning (NeSy 2025)

详情

展开后加载摘要…

URL PDF HTML 收藏
1807.01183 2018-08-22 cs.AI cs.LG cs.LO 66%

Quantified Markov Logic Networks

Víctor Gutiérrez-Basulto, Jean Christoph Jung, Ondrej Kuzelka

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

Comments Paper accepted at the 16th International Conference on Principles of Knowledge Representation and Reasoning (KR 2018). This work was also presented in the Eighth International Workshop on Statistical Relational AI (StarAI 2018) under the title "Markov Logic Networks with Statistical Quantifiers"

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.04177 2026-04-28 cs.CL 65%

Position: Logical Soundness is not a Reliable Criterion for Neurosymbolic Fact-Checking with LLMs

位置:逻辑正确性不是LLM神经符号事实核查中的可靠标准

Jason Chan, Robert Gaizauskas, Zhixue Zhao

机构 * School of Computer Science University of Sheffield(计算机科学学院 伦敦大学谢菲尔德分校)

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

AI总结 本文指出,逻辑正确性无法有效检测误导性声明,提出利用LLM的人类推理倾向来验证神经符号系统中的正式组件输出。

Comments ICLR 2026 Workshop on Logical Reasoning of Large Language Models

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.21364 2026-08-25 cs.CL cs.AI cs.MM 新提交 62%

Distinguishing Revision and Delayed Elaboration in Incremental Narrative Interpretation

增量叙事理解中修订与延迟阐释的区分

Yi-Chun Chen

机构 * National Cheng Kung University(成功大学)

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

AI总结 研究区分增量叙事理解中的修订与延迟阐释两种更新算子,以视觉叙事为域验证结构化表征可支持二者,其结构差异对增量推理及混合符号-神经系统具重要意义。

Comments Accepted as a full paper for presentation at the 2026 Workshop on Computational Models of Narrative (CMN 2026). This preprint corresponds to the workshop version

详情

展开后加载摘要…

URL PDF HTML 收藏
2607.20507 2026-08-25 cs.AI cs.LG 版本更新 62%

CacheSpec: Finding the Sweet Spot for Small Models in Large Language Models

MiniCache:用于高效大语言模型推理的具有小模型接口的可重复使用程序缓存

Jingquan Chen, Jie Feng, Jinghua Piao, Shaogang Hu, Yong Li

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

AI总结 研究针对大语言模型推理成本高的问题,提出MiniCache框架,将PoT程序转换为参数化缓存对象,通过重用小模型进行语义变量提取和推测性起草,减少目标大语言模型调用,实验证明其能提升推理性能,平衡延迟、缓存重用和准确性。

Comments Accepted by EMNLP 2026, codes via https://github.com/chenjqQAQ/CacheSpec

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.17843 2026-08-19 cs.CL cs.AI 新提交 62%

Encoded but Not Actionable: Auditing the Decode-Generate-Steer Gap in Frozen LLMs for Geometric Constraints

已编码但不可执行:审计冻结大语言模型中几何约束的解码-生成-引导差距

Man Liang, Xinzhao Cheng, Faizan Wajid

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

AI总结 该研究以参数化CAD约束为测试平台,审计6个冻结LLMs的解码-生成-引导差距,发现几何关系可解码性与生成、引导能力存在差异,区分了编码与表达控制失败。

Comments 13 pages, 7 figures, 8 tables, including appendices

详情

展开后加载摘要…

URL PDF HTML 收藏
2607.07436 2026-08-19 cs.AI cs.CL cs.CR 版本更新 62%

The Blind Curator: How a Biased Judge Silently Disables Skill Retirement in Self-Evolving Agents

盲选策展人:有偏见的评判者如何在自我进化的智能体中悄然阻碍技能淘汰

Xing Zhang, Yanwei Cui, Guanghui Wang, Ziyuan Li, Wei Qiu, Bing Zhu, Peiyang He

机构 * AWS Generative AI Innovation Center(亚马逊云科技生成式人工智能创新中心) HSBC Holdings Plc., HSBC Technology Center, China(汇丰控股有限公司,汇丰科技中心,中国)

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

AI总结 研究自我进化智能体中,有偏见的评判者对技能淘汰的影响。通过损坏奖励分析等方法发现,“误判通过”偏差会导致技能淘汰机制失效,此为行为安全结果。还提出廉价审计可判断评判者是否越过阈值影响智能体。

Comments Published at COLM 2026 Workshop on Agent Behavior

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.18747 2026-08-19 cs.IR cs.AI cs.CC cs.CL cs.DB 版本更新 62%

The $\mathbf{P}$-Completeness of Inverted Index Traversal: On the Complexity of Evaluating Boolean Query DAGs

倒排索引遍历的$\mathbf{P}$-完备性:关于布尔查询DAG评估的复杂性

Amir Aavani

机构 * Apple Inc.(苹果公司)

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

AI总结 本文证明基于DAG的布尔查询评估问题是$\mathbf{P}$-完备的,并提出一种稀疏感知算法ComputePN,通过正负对偶表示和DAG记忆化,将评估时间严格限制在$O(|Q| \cdot |U_{\mathit{active}}|)$,避免指数爆炸和全量扫描。

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.10723 2026-08-18 cs.CV cs.AI cs.LG cs.MA 版本更新 62%

AgentMV: A State-Guided Multi-Agent Framework for Budget-Aware Music Video Generation

AllocMV:通过结构化持久状态实现音乐视频生成的最优资源分配

Huimin Wang, Chang Xia, Leilei Ouyang, Yongqi Kang, Yu Fu, Yuqi Ouyang

机构 * College of Computer Science, Sichuan University(四川大学计算机学院)

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

AI总结 AllocMV提出了一种分层框架,将音乐视频合成建模为多重选择背包问题,通过动态规划优化资源分配,确保跨镜头一致性并降低生成成本。

Comments ECCV 2026 AI4VA

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.15808 2026-08-17 q-bio.QM cs.AI cs.CV cs.LG 62%

Foundation Models in Biomedical Imaging: Turning Hype into Reality

生物医学影像中的基础模型:从 hype 到现实

Amgad Muneer, Kai Zhang, Ibraheem Hamdi, Rizwan Qureshi, Muhammad Waqas, Shereen Fouad, Hazrat Ali, Syed Muhammad Anwar, Jia Wu

机构 * Department of Imaging Physics, The University of Texas MD Anderson Cancer Center(影像物理系,德克萨斯大学MD安德森癌症中心) Center for Secure Artificial Intelligence for Healthcare (SAFE), McWilliams School of Biomedical Informatics, UTHealth Houston(安全人工智能用于医疗保健中心(SAFE),麦威廉斯生物医学信息学学院,UTHealth休斯顿) Female Medicine in Machine Learning, Massachusetts Institute of Technology(机器学习中的女性医学,麻省理工学院) Department of Computer Science, Salim Habib University(计算机科学系,Salim Habib大学) School of Computer Science and Digital Technologies, Aston Centre for Artificial Intelligence Research and Application, Aston University(计算机科学与数字技术学院,阿斯顿人工智能研究与应用中心,阿斯顿大学) Division of Computing Science and Mathematics, University of Stirling(计算科学与数学系,斯特灵大学) School of Medicine and Health Sciences, George Washington University(医学与健康科学学院,乔治·华盛顿大学) Sheikh Zayed Institute for Pediatric Surgical Innovation, Children’s National Hospital(谢赫扎耶德儿童外科创新研究所,儿童医院) Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center(胸腔/头颈医学肿瘤科,德克萨斯大学MD安德森癌症中心)

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

AI总结 本文探讨了基础模型在生物医学影像中的应用,提出REAL-FM框架以评估模型的实际临床价值,指出基础模型在因果推理和安全性方面存在不足,强调需要协调的专业AI系统。

Comments 9 figures and 3 tables

Journal ref Nature Biomedical Engineering 10, 1557-1575 (2026)

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.12599 2026-08-14 cs.AI cs.LG 新提交 62%

Dead text or binding clause? Measuring and restoring constraint influence in black-box LLM dialogues

无效文本还是绑定条款?测量并恢复黑盒大语言模型对话中的约束影响

Haoyuan Zhu

机构 * University of Sheffield(谢菲尔德大学)

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

AI总结 该研究针对黑盒大语言模型对话中撤销约束失效的问题,提出 sysname 方法,通过合约账本、顺序消融探针和修复阶梯实现复发率的测量、预测与修复,降低了约束撤销的失效概率。

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.08849 2026-08-12 cs.CL cs.AI cs.DB cs.MA cs.SC 版本更新 62%

SatIR: Scalable High-Recall Constraint-Satisfaction-Based Information Retrieval for Clinical Trials Matching

SatIR:可扩展的高召回率约束满足基于信息检索的临床试验匹配

Zikai Zhou, Yufei Jin, Yilin Xu, Yu-Chiang Wang, Chieh-Ju Chao, Monica S. Lam

机构 * Department of Computer Science, Stanford University(斯坦福大学计算机科学系) Samueli Electrical and Computer Engineering, UCLA(UCLA Samueli电气与计算机工程系) Department of Computer Science and Informatics, Emory University(埃默里大学计算机科学与信息学系) Mayo Clinic(梅奥诊所)

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

AI总结 SatIR通过将临床试验资格条件和摘要转化为形式约束,结合SMT、关系代数和大语言模型,提升了临床试验匹配的召回率和效率,优于基于相似度的基线方法。

详情

展开后加载摘要…

URL PDF HTML 收藏
2409.03381 2026-08-12 cs.CL cs.AI 62%

CogniDual Framework: Self-Training Large Language Models within a Dual-System Theoretical Framework for Improving Cognitive Tasks

Yongxin Deng, Xihe Qiu, Xiaoyu Tan, Chao Qu, Jing Pan, Yuan Cheng, Yinghui Xu, Wei Chu

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

Journal ref IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2025), pp. 1-5

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.08118 2026-08-11 cs.AI cs.LG cs.SC math.CO 新提交 62%

Neurosymbolic Discovery of Algebraic Graph Constructions

代数图构造的神经符号发现

David Seka, Stefan Szeider

机构 * TU Wien(维也纳技术大学)

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

AI总结 针对仅提供原始图数据无法揭示其结构性质的问题,提出基于通用大语言模型与SageMath的神经符号智能体,可自动发现代数图构造,在100个高度对称图基准上全部成功,还找到Bernhart-Kainen可散度猜想的最小反例。

详情

展开后加载摘要…

URL PDF HTML 收藏
2508.09125 2026-08-11 cs.CL cs.LG 版本更新 62%

LogicIF: Towards Complex Logic Instruction Following

复杂逻辑指令生成

Mian Zhang, Shujian Liu, Sixun Dong, Ming Yin, Yebowen Hu, Xun Wang, Simin Ma, Song Wang, Sathish Reddy Indurthi, Haoyun Deng, Zhiyu Zoey Chen, Kaiqiang Song

机构 * University of Texas at Dallas(德克萨斯大学达拉斯分校) Independent Researcher(独立研究者) Duke University(杜克大学) University of Central Florida(佛罗里达大学中央分校)

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

AI总结 本文提出LogicIFGen和LogicIFEval,用于生成和评估复杂逻辑指令,揭示了当前LLMs在复杂指令跟随上的不足。

Comments COLM 2026 Acceptance

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.21724 2026-08-07 cs.LG cs.AI 62%

FISformer: Replacing Self-Attention with a Fuzzy Inference System in Transformer Models for Time Series Forecasting

FISformer:在时间序列预测的Transformer模型中用模糊推理系统替代自注意力

Bulent Haznedar, Levent Karacan

机构 * Computer Engineering Department, Gaziantep University(加扎莱普大学计算机工程系)

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

AI总结 FISFormer通过模糊推理机制替代传统注意力机制,提升时间序列预测中对不确定性和非线性依赖的建模能力,实验显示其在准确性、鲁棒性和可解释性上优于现有Transformer变体。

详情

展开后加载摘要…

URL PDF HTML 收藏