31 citations
- University of Minnesota, DuluthUS3 papers
- Mahidol UniversityTH2 papers
- Pennsylvania State UniversityUS2 papers
- Astronomical Institute of the Slovak Academy of SciencesSK1 paper
- Centre National de la Recherche ScientifiqueFR1 paper
- Concordia UniversityCA1 paper
- Georgia Southern UniversityUS1 paper
- Hodges UniversityUS1 paper
- Huawei Technologies (Canada)CA1 paper
- John Regan Associates (United Kingdom)GB1 paper
- Laboratoire de Physique Théorique et Modèles StatistiquesFR1 paper
- Michigan Technological UniversityUS1 paper
22 papers · 1 filter
Independent Patch Verification for Coding Agents with a Bidirectional Reconstruct-and-Verify Framework
Chenglin Li, Yisen Xu, Zehao Wang +3
Autonomous coding agents powered by large language models can now generate code patches directly from bug reports, but a fundamental gap remains: once a patch is produced, no mecha…
Turning Interaction History into Execution State: A Runtime Layer for Long-Horizon Coding Agents
Zehao Wang, Yisen Xu, Chenglin Li +5
Long-horizon coding agents accumulate hundreds of actions and observations in their trajectories, yet nothing in this record indicates which observations still describe the reposit…
Retrieval-Oriented Code Representations in Agentic Bug Localization
Genevieve Caumartin, Tse-Hsun, Chen +1
LLM-based agents are increasingly being used to support software development, yet their performance in repository-level tasks depends on retrieving the right code context. Existing…
Bug Report Specification Refinement with Trajectory Guidance for Automated Program Repair
S M Farah Al Fahim, Md Nakhla Rafi, Md Ahasanuzzaman +5
Bug reports serve as task specifications for repository-level automated program repair (APR) agents, but they often describe only the observed failure and omit repair-relevant info…
Rethinking Code Performance Benchmarks for LLMs
Nhat Minh Le, Yisen Xu, Zhijie Wang +2
Many function-level performance benchmarks have been proposed to evaluate whether large language models (LLMs) can generate efficient programs. However, results on these benchmarks…
CI-Repair-Bench: A Repository-Aware Benchmark for Automated Patch Validation via CI Workflows
Rabeya Khatun Muna, Md Nakhla Rafi, Tse-Hsun +1
Continuous Integration (CI) enforces repository-level correctness through multi-stage workflows and is central to modern software development, yet diagnosing and repairing CI failu…