HyperAgent：可规模化解决编码任务的通才软件工程智能体
HyperAgent: Generalist Software Engineering Agents to Solve Coding Tasks at Scale
AI总结：
针对现有自主软件智能体多仅适配特定软件工程功能的问题，提出含四个专用智能体的通才多智能体系统HyperAgent，模拟人类开发流程处理多语言全周期编码任务，在多个基准上超越现有最优水平。
摘要：
大语言模型（LLMs）已革新软件工程（SE）领域，在各类编码任务中展现出卓越能力。尽管近期进展已催生利用LLM完成端到端开发任务的自主软件智能体，但这类系统通常仅针对特定软件工程功能设计。我们提出HyperAgent，这是一种创新的通才多智能体系统，通过模拟人类开发者的工作流，可处理跨多种编程语言的广泛软件工程任务。HyperAgent包含四个专用智能体——规划器（Planner）、导航器（Navigator）、代码编辑器（Code Editor）和执行器（Executor），能够覆盖软件工程任务从初始规划到最终验证的全生命周期。HyperAgent在多项不同的软件工程任务中刷新了基准，包括在知名基准SWE-Bench上的GitHub问题解决任务，表现优于强基线。此外，HyperAgent在仓库级代码生成（RepoExec）以及故障定位与程序修复（Defects4J）任务中也展现出优异性能，通常超越现有最优基线。
英文摘要：
Large Language Models (LLMs) have revolutionized software engineering (SE), showcasing remarkable proficiency in various coding tasks. Despite recent advancements that have enabled the creation of autonomous software agents utilizing LLMs for end-to-end development tasks, these systems are typically designed for specific SE functions. We introduce HyperAgent, an innovative generalist multi-agent system designed to tackle a wide range of SE tasks across different programming languages by mimicking the workflows of human developers. HyperAgent features four specialized agents-Planner, Navigator, Code Editor, and Executor-capable of handling the entire lifecycle of SE tasks, from initial planning to final verification. HyperAgent sets new benchmarks in diverse SE tasks, including GitHub issue resolution on the renowned SWE-Bench benchmark, outperforming robust baselines. Furthermore, HyperAgent demonstrates exceptional performance in repository-level code generation (RepoExec) and fault localization and program repair (Defects4J), often surpassing state-of-the-art baselines.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
