代码大模型 / AI 编程

On the Use of Commit Messages for Corrective Software Maintenance: A Systematic Mapping Study

关于使用提交信息进行纠正性软件维护：一项系统映射研究

Syful Islam, Stefano Zacchiroli

专题命中 程序修复 ：program repair(abstract)；repository(abstract)；分类 cs.SE

AI总结 本文通过系统映射研究97篇2004至2025年发表的文献，探讨提交信息在纠正性维护中的应用，发现提交信息在 bug 分析和修复识别中作用显著，但其他领域如自动程序修复研究较少。

Comments Preprint. Accepted for publication at EASE 2026 (Track: Research Papers)

Learned or Memorized ? Quantifying Memorization Advantage in Code LLMs

学习还是记忆？量化代码大语言模型中的记忆优势

Djiré Albérick Euraste, Kaboré Abdoul Kader, Jordan Samhi, Earl T. Barr, Jacques Klein, Tegawendé F. Bissyandé

专题命中 程序修复 ：code generation(abstract)；program repair(abstract)；分类 cs.SE

AI总结 本文通过扰动方法量化代码LLM的记忆优势，评估8个开源代码LLM在19个基准测试中的表现，发现任务和模型差异显著，部分基准测试显示低敏感度，挑战了数据泄露的常见担忧。

Comments 12 pages, 12 figures. Accepted at ICSE 2026 (to appear)

NeuroProlog: Multi-Task Fine-Tuning for Neurosymbolic Mathematical Reasoning via the Cocktail Effect

NeuroProlog：通过鸡尾酒效应实现神经符号数学推理的多任务微调

Pratibha Zunjare, Michael Hsiao

机构 * Virginia Tech（弗吉尼亚理工大学） 纠错

专题命中 程序修复 ：program repair(abstract)；program synthesis(abstract)；分类 cs.AI

AI总结 NeuroProlog通过鸡尾酒效应实现神经符号数学推理的多任务微调，提升数学推理准确率并改进程序修复能力。

R3A: Reliable RTL Repair Framework with Multi-Agent Fault Localization and Stochastic Tree-of-Thoughts Patch Generation

R3A：一种可靠的RTL修复框架，结合多智能体故障定位和随机树状思维补丁生成

Zizhang Luo, Fan Cui, Kexing Zhou, Runlin Guo, Mile Xia, Hongyuan Hou, Yun Liang

机构 * Peking University（北京大学） ； Beihang University（北航） 纠错

专题命中 程序修复 ：program repair(abstract)；program synthesis(abstract)；分类 cs.AI

AI总结 R3A通过多智能体故障定位和随机树状思维补丁生成方法，有效提高了RTL修复的可靠性，修复率高达90.6%。

MORepair: Teaching LLMs to Repair Code via Multi-Objective Fine-tuning

Boyang Yang, Haoye Tian, Jiadong Ren, Hongyu Zhang, Jacques Klein, Tegawendé F. Bissyandé, Claire Le Goues, Shunfu Jin

专题命中 程序修复 ：program repair(abstract)；repository(abstract)；分类 cs.SE

Journal ref ACM Transactions on Software Engineering and Methodology, 2025

CodeReasoner: Enhancing the Code Reasoning Ability with Reinforcement Learning

Lingxiao Tang, He Ye, Zhongxin Liu, Xiaoxue Ren, Lingfeng Bao

专题命中 程序修复 ：code generation(abstract)；program repair(abstract)；分类 cs.SE

Extracting Fix Ingredients using Language Models

Julian Aron Prenner, Romain Robbes

专题命中 程序修复 ：code model(abstract)；program repair(abstract)；分类 cs.SE

NExT: Teaching Large Language Models to Reason about Code Execution

Ansong Ni, Miltiadis Allamanis, Arman Cohan, Yinlin Deng, Kensen Shi, Charles Sutton, Pengcheng Yin

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.CL、cs.LG

Comments 35 pages

A Study of Vulnerability Repair in JavaScript Programs with Large Language Models

Tan Khang Le, Saba Alimadadi, Steven Y. Ko

专题命中 程序修复 ：code generation(abstract)；program repair(abstract)；分类 cs.AI

Comments camera-ready version accepted to the short paper track at WWW'24

Large Language Models in Fault Localisation

Yonghao Wu, Zheng Li, Jie M. Zhang, Mike Papadakis, Mark Harman, Yong Liu

专题命中 程序修复 ：code generation(abstract)；program repair(abstract)；分类 cs.SE

RegMiner: Towards Constructing a Large Regression Dataset from Code Evolution History

Xuezhi Song, Yun Lin, Siang Hwee Ng, Yijian Wu, Xin Peng, Jin Song Dong, Hong Mei

专题命中 程序修复 ：program repair(abstract)；repository(abstract)；分类 cs.SE

Comments ISSTA'22

Ultra-Large Repair Search Space with Automatically Mined Templates: the Cardumen Mode of Astor

Matias Martinez, Martin Monperrus

专题命中 程序修复 ：program repair(abstract)；repository(abstract)；分类 cs.SE

Journal ref Proceedings of the International Symposium on Search Based Software Engineering, 2018

Extracting Concise Bug-Fixing Patches from Human-Written Patches in Version Control Systems

Yanjie Jiang, Hui Liu, Nan Niu, Lu Zhang, Yamin Hu

专题命中 程序修复 ：program repair(abstract)；repository(abstract)；分类 cs.SE

Comments ICSE 2021

ContraFix: Skill-Enhanced Contrastive Runtime Analysis for Vulnerability Repair

ContraFix：通过差分运行时证据和技能重用进行代理漏洞修复

Simiao Liu, Fang Liu, Peiding Wang, Taichuan Li, Yinghao Zhu, Xiaoli Lian, Li Zhang

机构 * Beihang University（北京航空航天大学） ； The University of Hong Kong（香港大学） 纠错

专题命中 程序修复 ：repository(abstract)；分类 cs.SE、cs.CL、cs.AI

AI总结 本文提出ContraFix框架，通过差分运行时证据和可重用的修复技能，解决大型语言模型代理在自动漏洞修复中的语义误解问题，实现了在SEC-Bench和PatchEval上的高准确率。

Comments Code: https://figshare.com/s/f173c78e44bca88ebaea

DiTOX: Fault Detection and Localization in the ONNX Optimizer

DiTOX：ONNX优化器中的故障检测与定位

Nikolaos Louloudakis, Ajitha Rajan

机构 * University of Edinburgh（爱丁堡大学） 纠错

专题命中 程序修复 ：repository(abstract)；分类 cs.SE、cs.AI、cs.LG

AI总结 DiTOX通过差分测试和故障定位技术，发现ONNX优化器中15个问题，揭示了优化器在不同任务中的正确性问题。

Comments 13 pages, 2 figures, 4 tables

LLM Chemistry Estimation for Multi-LLM Recommendation

Huascar Sanchez, Briland Hitaj

机构 * Computer Science Laboratory（计算机科学实验室） ； SRI International（SRI国际） 纠错

专题命中 程序修复 ：program repair(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 20 pages, 5 figures, 5 tables

Order Matters! An Empirical Study on Large Language Models' Input Order Bias in Software Fault Localization

Md Nakhla Rafi, Dong Jae Kim, Tse-Hsun Chen, Shaowei Wang

机构 * Software Performance, Analysis, \ Reliability (SPEAR) Lab\ University Montreal Canada ； DePaul University Chicago USA ； University of Manitoba Winnipeg Canada ； Software Performance, Analysis, \ Reliability (SPEAR) Lab\ University ； DePaul University ； University of Manitoba 纠错

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.LG

InvAASTCluster: On Applying Invariant-Based Program Clustering to Introductory Programming Assignments

Pedro Orvalho, Mikoláš Janota, Vasco Manquinho

机构 * INESC-ID ； Instituto Superior Técnico（理工学院） ； Universidade de Lisboa（里斯本大学） ； CIIRC ； Czech Technical University in Prague（布拉格捷克技术大学） 纠错

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.PL

Comments 31 pages, 21 Figures, 5 Tables. Accepted for publication at the Journal of Systems and Software. GitHub repo: https://github.com/pmorvalho/InvAASTCluster

Fully Autonomous Programming using Iterative Multi-Agent Debugging with Large Language Models

Anastasiia Grishina, Vadim Liventsev, Aki Härmä, Leon Moonen

专题命中 程序修复 ：program synthesis(abstract)；分类 cs.SE、cs.AI、cs.LG

Comments Accepted for publication in ACM Trans. Evol. Learn. Optim., February 2025. arXiv admin note: text overlap with arXiv:2304.10423

Effective Large Language Model Debugging with Best-first Tree Search

Jialin Song, Jonathan Raiman, Bryan Catanzaro

专题命中 程序修复 ：code generation(abstract)；分类 cs.SE、cs.AI、cs.LG

DebugBench: Evaluating Debugging Capability of Large Language Models

Runchu Tian, Yining Ye, Yujia Qin, Xin Cong, Yankai Lin, Yinxu Pan, Yesai Wu, Haotian Hui, Weichuan Liu, Zhiyuan Liu, Maosong Sun

专题命中 程序修复 ：code generation(abstract)；分类 cs.SE、cs.CL、cs.AI

Comments Accepted as Findings of ACL 2024

Automatic Programming: Large Language Models and Beyond

Michael R. Lyu, Baishakhi Ray, Abhik Roychoudhury, Shin Hwei Tan, Patanamon Thongtanunam

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.LG

DeepCode AI Fix: Fixing Security Vulnerabilities with Large Language Models

Berkay Berabi, Alexey Gronskiy, Veselin Raychev, Gishor Sivanrupan, Victor Chibotaru, Martin Vechev

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.LG、cs.PL

Comments 26 pages, 13 figures (v2, small fix in author affiliations)

Enhancing Genetic Improvement Mutations Using Large Language Models

Alexander E. I. Brownlee, James Callan, Karine Even-Mendoza, Alina Geiger, Carol Hanna, Justyna Petke, Federica Sarro, Dominik Sobania

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.LG

Comments Accepted for publication at the Symposium on Search-Based Software Engineering (SSBSE) 2023

Journal ref Arcaini, P., Yue, T., Fredericks, E.M. (eds) Search-Based Software Engineering. SSBSE 2023. Lecture Notes in Computer Science, vol 14415. Springer, Cham

Zipr: A High-Impact, Robust, Open-source, Multi-platform, Static Binary Rewriter

Jason D. Hiser, Anh Nguyen-Tuong, Jack W. Davidson

专题命中 程序修复 ：program repair(abstract)；repository(abstract)

Comments 5 pages

Graph Neural Networks For Mapping Variables Between Programs -- Extended Version

Pedro Orvalho, Jelle Piepenbrock, Mikoláš Janota, Vasco Manquinho

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.LG

Comments Extended version of "Graph Neural Networks For Mapping Variables Between Programs", paper accepted at ECAI 2023. Github: https://github.com/pmorvalho/ecai23-GNNs-for-mapping-variables-between-programs. 11 pages, 5 figures, 4 tables and 3 listings

SampleFix: Learning to Generate Functionally Diverse Fixes

Hossein Hajipour, Apratim Bhattacharyya, Cristian-Alexandru Staicu, Mario Fritz

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.LG、cs.PL

Comments 15 pages, 6 figures

Patching as Translation: the Data and the Metaphor

Yangruibo Ding, Baishakhi Ray, Premkumar Devanbu, Vincent J. Hellendoorn

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.LG、cs.PL

Learning to Fix Build Errors with Graph2Diff Neural Networks

Daniel Tarlow, Subhodeep Moitra, Andrew Rice, Zimin Chen, Pierre-Antoine Manzagol, Charles Sutton, Edward Aftandilian

专题命中 程序修复 ：program repair(abstract)；分类 cs.SE、cs.AI、cs.LG

Comments Submitted for review on Aug 23, 2019

A Comprehensive Evaluation of Code Language Models for Security Patch Detection

以代码为中心的漏洞修复提交检测：一个统一的基准和实证研究

Nils Loose, Joseph Bienhüls, Kristoffer Hempel, Felix Mächtle, Thomas Eisenbarth

机构 * University of Lübeck（吕贝克大学） ； University of Lübeck Institute for IT Security（吕贝克大学信息安全部） 纠错

专题命中 程序修复 ：repository(abstract)；分类 cs.SE、cs.LG

AI总结 本文通过统一框架评估了基于代码语言模型的漏洞修复提交检测，发现仅依赖代码变更无法让模型获得可转移的安全相关代码理解，且在去除提交信息后，增加内过程语义上下文也无法使模型关注代码变更，性能受数据分布和时间分割影响显著。

Comments Accepted at ASE26