代码大模型 / AI 编程

Beyond Attention Heatmaps: How to Get Better Explanations for Multiple Instance Learning Models in Histopathology

超越注意力热图：如何为多实例学习模型在病理学中的更好解释

Mina Jamshidi Idaji, Julius Hense, Tom Neuhäuser, Augustin Krause, Yanqing Luo, Oliver Eberle, Thomas Schnake, Laure Ciernik, Farnoush Rezaei Jafari, Reza Vahidimajd, Jonas Dippel, Christoph Walz, Frederick Klauschen, Andreas Mock, Klaus-Robert Müller

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

AI总结 本研究提出了一种评估MIL热图质量的通用框架，发现扰动、LRP和IG在病理学任务中表现更优，展示了其在生物验证和发现不同模型策略中的应用。

What Papers Don't Tell You: Recovering Tacit Knowledge for Automated Paper Reproduction

论文未告诉你什么：为自动化论文复现恢复隐性知识

Lehui Li, Ruining Wang, Haochen Song, Yaoxin Mao, Tong Zhang, Yuyao Wang, Jiayi Fan, Yitong Zhang, Jieping Ye, Chengqi Zhang, Yongshun Gong

机构 * School of Software, Shandong University China ； Beijing Institute of Technology China ； Zhejiang University China ； Dept.\ of Math \& Statistics, Boston University USA ； College of AI, Tsinghua University China ； Alibaba Group China ； The Hong Kong Polytechnic University Hong Kong SAR, China ； School of Software, Shandong University ； Beijing Institute of Technology ； Zhejiang University ； Dept.\ of Math \& Statistics, Boston University ； College of AI, Tsinghua University ； Alibaba Group ； The Hong Kong Polytechnic University 纠错

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

AI总结 本文提出一种基于图的智能体框架，通过恢复关系型、身体型和集体型隐性知识，提升自动化论文复现性能。

Comments 32 pages (+ appendix), 8 figures. Lehui Li and Ruining Wang contributed equally. Yongshun Gong is the corresponding author

Historian: Reducing Manual Validation in APR Benchmarking via Evidence-Based Assessment

Historian: 通过基于证据的评估减少APR基准测试中的手动验证

Sahand Moslemi, Mayasah Lami, Anil Koyuncu

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

AI总结 Historian通过基于证据的多参考比较，提高自动化程序修复评估的准确性与覆盖率，减少手动验证需求。

HAFix: History-Augmented Large Language Models for Bug Fixing

HAFix: 历史增强型大语言模型用于缺陷修复

Yu Shi, Abdul Ali Bangash, Emad Fallahzadeh, Bram Adams, Ahmed E. Hassan

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

AI总结 HAFix通过整合历史启发式方法提升LLM在缺陷修复中的性能，实验表明其在缺陷修复率上显著优于基线方法。

Comments Accepted at Empirical Software Engineering (EMSE) 2026 Journal

CloudFix: Automated Policy Repair for Cloud Access Control Policies Using Large Language Models

CloudFix: 利用大语言模型实现云访问控制策略的自动化修复

Bethel Hall, Owen Ungaro, William Eiers

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

AI总结 CloudFix利用大语言模型和形式方法，实现云访问控制策略的自动化修复，提高修复准确性和效率。

Comments 12 pages

Journal ref SANER 2026

Multi-Agent End-to-End Vulnerability Management for Mitigating Recurring Vulnerabilities

多智能体端到端漏洞管理用于缓解重复漏洞

Zelong Zheng, Jiayuan Zhou, Xing Hu, Yi Gao, Shengyi Pan

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

AI总结 MAVM通过多智能体框架实现端到端重复漏洞管理，整合知识库、检测、确认、修复和验证，有效提升漏洞修复准确率。

Leveraging Mutation Analysis for LLM-based Repair of Quantum Programs

利用突变分析改进基于大语言模型的量子程序修复

Chihiro Yoshida, Yuta Ishimoto, Olivier Nourry, Masanari Kondo, Makoto Matsushita, Yasutaka Kamei, Yoshiki Higo

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

AI总结 本文提出利用突变分析提升基于大语言模型的量子程序修复效果，通过设计不同提示配置实验，发现突变分析能显著提高修复成功率和解释质量。

Comments 6 pages, Accepted at SANER-ERA 2026

TwinFormer: A Dual-Level Transformer for Long-Sequence Time-Series Forecasting

TwinFormer：一种用于长序列时间序列预测的双层Transformer

Mahima Kumavat, Aditya Maheshwari

机构 * Operations Management and Quantitative Techniques Area, Indian Institute of Management Indore（印度管理学院印第安纳分校运营管理与定量技术部门） 纠错

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

AI总结 TwinFormer通过双阶段Transformer架构，结合top-k稀疏注意力和GRU聚合，实现高效长序列时间序列预测，优于多种现有方法。

Comments 14 pages, 4 figures

Parameter-Efficient Fine-Tuning with Attributed Patch Semantic Graph for Automated Patch Correctness Assessment

基于属性补丁语义图的参数高效微调用于自动补丁正确性评估

Zhenyu Yang, Jingwen Wu, Zhen Yang, Zhongxing Yu

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

AI总结 本文提出基于属性补丁语义图的参数高效微调方法，用于提升自动补丁正确性评估的准确性和F1分数。

Comments 20 pages, 5 figures, 18 tables

Statistical Independence Aware Caching for LLM Workflows

面向LLM工作流的统计独立性感知缓存

Yihan Dai, Dimitrios Stamatios Bouras, Haoxiang Jia, Sergey Mechtaev

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

AI总结 Mnimi通过封装统计约束在LLM引用类型中，提升代码修复工作流的可重复性、调试效率和统计正确性。

Attention Pruning: Automated Fairness Repair of Language Models via Surrogate Simulated Annealing

注意力剪枝：通过代理模拟退火实现语言模型的自动公平性修复

Vishnu Asutosh Dasu, Md Rafi ur Rashid, Vipul Gupta, Saeid Tizpaz-Niari, Gang Tan

机构 * Pennsylvania State University（宾夕法尼亚州立大学） ； University of Illinois（伊利诺伊大学） 纠错

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

AI总结 本文提出注意力剪枝方法，通过代理模拟退火在LLMs中减少偏见，提升公平性。

Towards a Human-in-the-Loop Framework for Reliable Patch Evaluation Using an LLM-as-a-Judge

Sherry Shi, Renyao Wei, Michele Tufano, José Cambronero, Runxiang Cheng, Franjo Ivančić, Pat Rondon

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

Designing for Novice Debuggers: A Pilot Study on an AI-Assisted Debugging Tool

Oka Kurniawan, Erick Chandra, Christopher M. Poskitt, Yannic Noller, Kenny Tsu Wei Choo, Cyrille Jegourel

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

Comments Accepted by the 25th Koli Calling International Conference on Computing Education Research (Koli Calling 2025)

Journal ref Proc. Koli Calling '25, Article No. 41, pages 1-7. ACM, 2025

Do Automated Fixes Truly Mitigate Smart Contract Exploits?

Sofia Bobadilla, Monica Jin, Martin Monperrus

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

Journal ref IEEE Transactions on Software Engineering, 2025

InspectCoder: Dynamic Analysis-Enabled Self Repair through interactive LLM-Debugger Collaboration

Yunkun Wang, Yue Zhang, Guochang Li, Chen Zhi, Binhua Li, Fei Huang, Yongbin Li, Shuiguang Deng

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

MAHL: Multi-Agent LLM-Guided Hierarchical Chiplet Design with Adaptive Debugging

Jinwei Tang, Jiayin Qin, Nuo Xu, Pragnya Sudershan Nalla, Yu Cao, Yang, Zhao, Caiwen Ding

机构 * University of Minnesota - Twin Cities（明尼苏达大学-双城分校） 纠错

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

Automated Repair of OpenID Connect Programs (Extended Version)

Tamjid Al Rahat, Yanju Chen, Yu Feng, Yuan Tian

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

Comments This is an extended version. The original paper is accepted to ASE 2025

Analyzing the Evolution and Maintenance of Quantum Software Repositories

Krishna Upadhyay, Vinaik Chhetri, A. B. Siddique, Umar Farooq

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

Comments 12 pages, 12 figures, 6 tables,

Journal ref IEEE Quantum Software Development (QSW), 2025

Crash Report Enhancement with Large Language Models: An Empirical Study

S M Farah Al Fahim, Md Nakhla Rafi, Zeyang Ma, Dong Jae Kim, Tse-Hsun, Chen

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

Integrating Large Language Models in Software Engineering Education: A Pilot Study through GitHub Repositories Mining

Maryam Khan, Muhammad Azeem Akbar, Jussi Kasurinen

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

Mokav: Execution-driven Differential Testing with LLMs

Khashayar Etemadi, Bardia Mohammadi, Zhendong Su, Martin Monperrus

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

Journal ref Journal of Systems and Software, 2025

LLMCup: Ranking-Enhanced Comment Updating with LLMs

Hua Ge, Juan Zhai, Minxue Pan, Fusen He, Ziyue Tan

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

Comments 13 pages, 10 figures

Focus Your Attention: Towards Data-Intuitive Lightweight Vision Transformers

Suyash Gaurav, Muhammad Farhan Humayun, Jukka Heikkonen, Jatin Chaudhary

机构 * Tokyo International University Tokyo（东京国际大学） ； University of Turku（图尔库大学） 纠错

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

PAGENT: Learning to Patch Software Engineering Agents

Haoran Xue, Gias Uddin, Song Wang

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

Adversarial Reasoning for Repair Based on Inferred Program Intent

He Ye, Aidan Z. H. Yang, Chang Hu, Yanlin Wang, Tao Zhang, Claire Le Goues

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

Equational Anti-Unification over Absorption Theories

Mauricio Ayala-Rincon, David M. Cerna, Andres Felipe Gonzalez Barragan, Temur Kutsia

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

Comments 23 pages, 18 main text, under review

InfraFix: Technology-Agnostic Repair of Infrastructure as Code

Nuno Saavedra, João F. Ferreira, Alexandra Mendes

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

Comments Accepted at ISSTA 2025

An Anatomy of 488 Faults from Defects4J Based on the Control- and Data-Flow Graph Representations of Programs

Alexandra van der Spuy, Bernd Fischer

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

Comments 6 pages, 5 figures, EASE 2025 conference

RePurr: Automated Repair of Block-Based Learners' Programs

Sebastian Schweikl, Gordon Fraser

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

Comments 24 pages, ACM International Conference on the Foundations of Software Engineering (FSE 2025)

Agent That Debugs: Dynamic State-Guided Vulnerability Repair

Zhengyao Liu, Yunlong Ma, Jingxuan Xu, Junchen Ai, Xiang Gao, Hailong Sun, Abhik Roychoudhury

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