Survey maps LLM agents for software issue resolution

A new survey paper details the advancements in using large language models (LLMs) for agentic software issue resolution. The paper, authored by Zhonghao Jiang, systematically reviews 242 recent studies in this emerging field. It categorizes existing research across benchmarks, techniques, and empirical studies, highlighting reinforcement learning as a key training paradigm for these agentic systems in software engineering. The survey also identifies current challenges and proposes future research directions to further bridge artificial intelligence and software engineering. AI

IMPACT Provides a structured overview of LLM applications in software maintenance, guiding future research and development in agentic AI for engineering tasks.

RANK_REASON The item is a survey paper on a research topic. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Agentic Software Issue Resolution
• artificial intelligence
• arXiv
• large-language models
• reinforcement learning
• software engineering
• Zhonghao Jiang

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →