arXivDaily arXiv每日学术速递 周一至周五更新

AI 大模型

代码大模型 / AI 编程

代码生成、软件工程智能体、程序修复、测试生成和开发者工具。

共收录 189 信号源:cs.SE, cs.CL, cs.AI, cs.LG, cs.PL

1. 程序分析与验证 189 篇

2607.08339 2026-07-10 cs.CL cs.AI cs.PL 新提交 82%

TypeProbe: Recovering Type Representations from Hidden States of Pre-trained Code Models

TypeProbe:从预训练代码模型的隐藏状态中恢复类型表示

Giuliano Gorgone, Fausto Carcassi

机构 * ILLC, University of Amsterdam(逻辑、语言与计算研究所,阿姆斯特丹大学)

专题命中 程序分析与验证 :code model(title,abstract);分类 cs.CL、cs.AI、cs.PL

AI总结 研究代码模型内部编码类型信息的情况,利用Java和Python代码示例的并行数据集探测预训练代码模型的残差流获取类型表示,发现跨语言类型表示等结果,且结构对一些变化有鲁棒性,还发布了代码和数据集。

Comments 18 pages, 12 figures. Accepted at ESSLLI 2026 (StuS; double-blind)

详情

展开后加载摘要…

URL PDF HTML 收藏
2402.11842 2024-11-22 cs.SE cs.AI cs.CL 82%

CodeArt: Better Code Models by Attention Regularization When Symbols Are Lacking

Zian Su, Xiangzhe Xu, Ziyang Huang, Zhuo Zhang, Yapeng Ye, Jianjun Huang, Xiangyu Zhang

专题命中 程序分析与验证 :code model(title,abstract);分类 cs.SE、cs.CL、cs.AI

详情

展开后加载摘要…

URL PDF HTML 收藏
2409.09271 2024-09-17 cs.SE cs.PL 81%

Python Symbolic Execution with LLM-powered Code Generation

Wenhan Wang, Kaibo Liu, An Ran Chen, Ge Li, Zhi Jin, Gang Huang, Lei Ma

专题命中 程序分析与验证 :code generation(title,abstract);分类 cs.SE、cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏
2507.09665 2025-07-15 cs.SE 79%

Is Quantization a Deal-breaker? Empirical Insights from Large Code Models

Saima Afrin, Bowen Xu, Antonio Mastropaolo

专题命中 程序分析与验证 :code model(title,abstract);分类 cs.SE

详情

展开后加载摘要…

URL PDF HTML 收藏
1603.03533 2016-03-14 cs.PL 79%

Stateless Code Model Checking of Information Flow Security

Elaheh Ghassabani, Mohammad Abdollahi Azgomi

专题命中 程序分析与验证 :code model(title,abstract);分类 cs.PL

Comments Trustworthy Computing Laboratory, School of Computer Engineering, Iran University of Science and Technology, Tehran, Iran, 2014

详情

展开后加载摘要…

URL PDF HTML 收藏
1508.07829 2015-09-01 cs.LO cs.PL 79%

Using Program Synthesis for Program Analysis

Cristina David, Daniel Kroening, Matt Lewis

专题命中 程序分析与验证 :program synthesis(title,abstract);分类 cs.PL

Comments 19 pages, to appear in LPAR 2015. arXiv admin note: text overlap with arXiv:1409.4925

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.12409 2026-01-21 quant-ph hep-th math-ph math.MP math.QA 78%

Topological quantum color code model on infinite lattice

无限晶格上的拓扑量子颜色码模型

Shiyu Cao, Zhian Jia, Sheng Tan

专题命中 程序分析与验证 :code model(title,abstract)

AI总结 本研究在无限晶格中分析了颜色码模型,揭示其拓扑序由双层扭码描述,并通过构造弦算子和研究激发性质验证了其拓扑结构。

Comments v1: comments are welcome

详情

展开后加载摘要…

URL PDF HTML 收藏
2309.10037 2024-11-28 hep-th cond-mat.str-el quant-ph 78%

A stabilizer code model with non-invertible symmetries: Strange fractons, confinement, and non-commutative and non-Abelian fusion rules

Tanay Kibe, Ayan Mukhopadhyay, Pramod Padmanabhan

专题命中 程序分析与验证 :code model(title,abstract)

Comments 56 pages, 17 figures; v2: more clarifications about the properties of symmetry transformations added; v3: expanded discussion section; v4: one figure added and further discussion about symmetries included

Journal ref Physical Review D 110, 105019 (2024)

详情

展开后加载摘要…

URL PDF HTML 收藏
1910.10309 2020-03-11 cond-mat.stat-mech cond-mat.str-el math-ph math.MP quant-ph 78%

A dynamical Toric Code model with fusion and de-fusion

Bruno Nachtergaele, Nicholas E. Sherman

专题命中 程序分析与验证 :code model(title,abstract)

Journal ref Phys. Rev. B 101, 115105 (2020)

详情

展开后加载摘要…

URL PDF HTML 收藏
cond-mat/0608379 2009-12-01 cond-mat.dis-nn 78%

Generalized Rate-Code Model for Neuron Ensembles with Finite Populations

Hideo Hasegawa

专题命中 程序分析与验证 :code model(title,abstract)

Comments 19 pages, 9 figures, accepted in Phys. Rev. E after minor modifications

Journal ref Phys. Rev. E 75 (2007) 051904

详情

展开后加载摘要…

URL PDF HTML 收藏
2607.21491 2026-07-24 cs.CL cs.LG 新提交 76%

What, Where, and How: Disentangling the Roles of Task, Language, and Model in Code Model Representations

是什么、在哪里以及如何:解开任务、语言和模型在代码模型表示中的作用

Piotr Wilam

机构 * University College London(伦敦大学学院)

专题命中 程序分析与验证 :code model(title);分类 cs.CL、cs.LG

AI总结 研究独立训练的语言模型在代码表示上的情况,通过2x2设计扩展概念电路提取方法,测量语法概念,发现任务决定概念获专用电路,模型决定电路位置及增长方式,还得出如Rust与Python电路差异等结论,为后续测试奠定基础。

Comments 16 pages, 11 figures, 6 tables. Code: https://github.com/piotrwilam/Atlas2x2 ; dataset: https://huggingface.co/datasets/piotrwilam/Atlas2x2

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.07814 2025-12-10 cs.SE cs.AI cs.CR 76%

Understanding Privacy Risks in Code Models Through Training Dynamics: A Causal Approach

通过训练动态理解代码模型中的隐私风险:一种因果方法

Hua Yang, Alejandro Velasco, Sen Fang, Bowen Xu, Denys Poshyvanyk

机构 * North Carolina State University(北卡罗来纳州立大学) William & Mary(威廉与玛丽学院)

专题命中 程序分析与验证 :code model(title);分类 cs.SE、cs.AI

AI总结 本研究通过因果方法分析代码模型中不同PII类型的泄露风险差异,揭示了泄露风险与可学习性之间的因果关系,并为开发类型感知的防御策略提供依据。

Comments 21 pages, 8 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2306.10763 2023-11-06 cs.CL cs.AI cs.LG cs.PL cs.SE 74%

Guiding Language Models of Code with Global Context using Monitors

Lakshya A Agrawal, Aditya Kanade, Navin Goyal, Shuvendu K. Lahiri, Sriram K. Rajamani

专题命中 程序分析与验证 :repository(abstract,comments);分类 cs.SE、cs.CL、cs.AI

Comments Accepted to NeurIPS 2023 and to appear as "Monitor-Guided Decoding of Code LMs with Static Analysis of Repository Context" at https://neurips.cc/virtual/2023/poster/70362 . Contents: 11 pages, 15 additional pages of appendix, 13 figures, 3 tables

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.13927 2026-04-16 cs.PL 74%

AI Coding Agents Need Better Compiler Remarks

AI 编程代理需要更好的编译器注释

Akash Deo, Simone Campanoni, Tommy McMichen

专题命中 程序分析与验证 :coding agent(title);分类 cs.PL

AI总结 本文研究了编译器注释对AI代理优化程序的影响,发现精确注释能显著提升性能,而模糊注释会导致语义错误,结论是未来编译器需提供结构化反馈以支持自主性能工程。

Comments 3 pages, 1 figure, 2 tables, Presented at Workshop on Co-Design for Agentic and Multimodal AI (CoDAIM) 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.09995 2026-04-14 eess.SY cs.AI cs.SY 74%

Agentic Application in Power Grid Static Analysis: Automatic Code Generation and Error Correction

代理在电力网络静态分析中的应用:自动代码生成与错误修正

Qinjuan Wang, Shan Yang, Yongli Zhu

机构 * Sun Yat-sen University(中山大学)

专题命中 程序分析与验证 :code generation(title);分类 cs.AI

AI总结 本文提出一个基于LLM的代理,通过将自然语言转换为MATLAB脚本实现电力网络静态分析自动化,采用三重纠错系统提升代码准确性,实验表明其在复杂任务中能有效消除幻觉。

Comments This paper has been accepted for presentation at the 9th International Conference on Energy, Electrical and Power Engineering (CEEPE 2026) in Nanjing, China, April 17-19, 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2602.16981 2026-02-20 cs.PL 74%

Mason: Type- and Name-Guided Program Synthesis

Mason: 基于类型和名称的程序合成

Jasper Geer, Fox Huston, Jeffrey S. Foster

专题命中 程序分析与验证 :program synthesis(title);分类 cs.PL

AI总结 Mason通过类型和名称引导合成技术,自动将设计模式插入到面向对象程序中,提升了多类程序合成的效率和准确性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.00509 2026-01-05 cs.CR cs.LG 74%

Improving LLM-Assisted Secure Code Generation through Retrieval-Augmented-Generation and Multi-Tool Feedback

通过检索增强生成和多工具反馈改进LLM辅助的安全代码生成

Vidyut Sriram, Sawan Pandita, Achintya Lakshmanan, Aneesh Shamraj, Suman Saha

机构 * Pennsylvania State University(宾夕法尼亚州立大学)

专题命中 程序分析与验证 :code generation(title);分类 cs.LG

AI总结 通过检索增强生成和多工具反馈,改进LLM辅助的安全代码生成,显著减少安全漏洞和缺陷率。

详情

展开后加载摘要…

URL PDF HTML 收藏
2203.16697 2022-04-06 cs.PL 74%

Type-Directed Program Synthesis for RESTful APIs

Zheng Guo, David Cao, Davin Tjong, Jean Yang, Cole Schlesinger, Nadia Polikarpova

专题命中 程序分析与验证 :program synthesis(title);分类 cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏
1911.04091 2022-04-01 cs.PL 74%

Program Synthesis by Type-Guided Abstraction Refinement

Zheng Guo, Michael James, David Justo, Jiaxiao Zhou, Ziteng Wang, Ranjit Jhala, Nadia Polikarpova

专题命中 程序分析与验证 :program synthesis(title);分类 cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏
2202.03293 2022-02-08 cs.PL 74%

Composable and Modular Code Generation in MLIR: A Structured and Retargetable Approach to Tensor Compiler Construction

Nicolas Vasilache, Oleksandr Zinenko, Aart J. C. Bik, Mahesh Ravishankar, Thomas Raoux, Alexander Belyaev, Matthias Springer, Tobias Gysi, Diego Caballero, Stephan Herhut, Stella Laurenzo, Albert Cohen

专题命中 程序分析与验证 :code generation(title);分类 cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏
2104.12131 2021-04-27 cs.SE 74%

Exploring the links between software development task type, team attitudes and task completion performance: Insights from the Jazz repository

Sherlock A. Licorish, Stephen G. MacDonell

专题命中 程序分析与验证 :repository(title);分类 cs.SE

Comments Journal Paper, 19 pages, 3 figures, 8 tables

Journal ref Information and Software Technology 97(2018), pp.10-25

详情

展开后加载摘要…

URL PDF HTML 收藏
2102.13183 2021-04-09 cs.PL 74%

RbSyn: Type- and Effect-Guided Program Synthesis

Sankha Narayan Guria, Jeffrey S. Foster, David Van Horn

专题命中 程序分析与验证 :program synthesis(title);分类 cs.PL

Comments 17 pages, 13 figures, this is a technical report (with appendix) of a paper to appear in Programming Language Design and Implementation (PLDI) 2021

详情

展开后加载摘要…

URL PDF HTML 收藏
1206.6466 2012-07-02 cs.NE cs.MS cs.PL 74%

Utilizing Static Analysis and Code Generation to Accelerate Neural Networks

Lawrence McAfee, Kunle Olukotun

专题命中 程序分析与验证 :code generation(title);分类 cs.PL

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

详情

展开后加载摘要…

URL PDF HTML 收藏
2507.05269 2026-01-21 cs.SE cs.AI 73%

CoRe: Benchmarking LLMs Code Reasoning Capabilities through Static Analysis Tasks

CoRe:通过静态分析任务基准测试LLM的代码推理能力

Danning Xie, Mingwei Zheng, Xuwei Liu, Jiannan Wang, Chengpeng Wang, Lin Tan, Xiangyu Zhang

机构 * Department of Computer Science Purdue University(计算机科学系 帕克森大学)

专题命中 程序分析与验证 :code generation(abstract);program repair(abstract);分类 cs.SE、cs.AI

AI总结 CoRe通过静态分析任务评估LLM的代码推理能力,发现主流模型在深层次语义理解和多步骤推理上存在不足。

Comments NeurIPS 2025 Datasets & Benchmarks Spotlight

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.12466 2025-09-23 cs.LG cs.AI cs.CL cs.PL cs.SE 73%

EquiBench: Benchmarking Large Language Models' Reasoning about Program Semantics via Equivalence Checking

Anjiang Wei, Jiannan Cao, Ran Li, Hongyu Chen, Yuhui Zhang, Ziheng Wang, Yuan Liu, Thiago S. F. X. Teixeira, Diyi Yang, Ke Wang, Alex Aiken

机构 * Stanford University(斯坦福大学) MIT(麻省理工学院) Google(谷歌) Nanjing University(南京大学) Intel(英特尔)

专题命中 程序分析与验证 :code generation(abstract);分类 cs.SE、cs.CL、cs.AI

详情

展开后加载摘要…

URL PDF HTML 收藏
2405.17238 2025-04-08 cs.CR cs.PL cs.SE 73%

IRIS: LLM-Assisted Static Analysis for Detecting Security Vulnerabilities

Ziyang Li, Saikat Dutta, Mayur Naik

专题命中 程序分析与验证 :code generation(abstract);repository(abstract);分类 cs.SE、cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏
2210.07909 2025-07-22 cond-mat.str-el math-ph math.MP quant-ph 71%

Modified toric code models with flux attachment from Hopf algebra gauge theory

Mia Conlon, Domenico Pellegrino, J. K. Slingerland

专题命中 程序分析与验证 :code model(title)

Comments 37 pages, 15 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2607.09217 2026-07-13 cs.AI cs.MS 新提交 70%

OpenProver: Agentic and Interactive Theorem Proving with Lean 4

OpenProver:使用Lean 4进行智能且交互式的定理证明

Matěj Kripner, Milan Straka

机构 * Charles University, Faculty of Mathematics and Physics(查尔斯大学数学与物理系)

专题命中 程序分析与验证 :code generation(abstract);repository(abstract);分类 cs.AI

AI总结 介绍用于大语言模型驱动的自动化定理证明的开源系统OpenProver,它集成特定架构,完全开源,能自动验证证明,提供交互式界面,通过在ProofNet上评估展示自动验证在定量消融实验中的潜力。

Comments 7 pages, 2 figures. Accepted at the 19th Conference on Intelligent Computer Mathematics (CICM 2026)

详情

展开后加载摘要…

URL PDF HTML 收藏
2510.22907 2026-05-05 cs.CL cs.AI cs.PL cs.SE 70%

Reinforcement Learning from Compiler and Language Server Feedback

从编译器和语言服务器反馈中学习强化学习

Yifan Zhang, Lanser Contributors

机构 * Princeton University(普林斯顿大学)

专题命中 程序分析与验证 :coding agent(abstract);分类 cs.SE、cs.CL、cs.AI

AI总结 本文提出RLCSF方法和Lanser-CLI工具,通过编译器和语言服务器的反馈信号提升编码代理的可靠性,利用确定性奖励函数和分析捆绑技术实现过程监督。

Comments Project Page: https://github.com/yifanzhang-pro/lanser-cli

详情

展开后加载摘要…

URL PDF HTML 收藏
2404.18249 2024-12-17 cs.PL 70%

Tenspiler: A Verified Lifting-Based Compiler for Tensor Operations (Extended Version)

Jie Qiu, Colin Cai, Sahil Bhatia, Niranjan Hasabnis, Sanjit A. Seshia, Alvin Cheung

专题命中 程序分析与验证 :code generation(abstract);program synthesis(abstract);分类 cs.PL

详情

展开后加载摘要…

URL PDF HTML 收藏