arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.07927cs.SEcs.AI
收藏

SWE-Fuse: 通过无问题轨迹学习和熵感知RLVR训练增强软件代理

SWE-Fuse: Empowering Software Agents via Issue-free Trajectory Learning and Entropy-aware RLVR Training

Ant Group(蚂蚁集团);The Chinese University of Hong Kong(香港中文大学)

Xin-Cheng Wen, Binbin Chen, Haoxuan Lan, Hang Yu, Peng Di, Cuiyun Gao

AI总结:

SWE-Fuse通过融合问题描述引导和无问题样本,提升软件代理在解决现实问题中的性能,其熵感知训练模块有效平衡探索与稳定性。

摘要:

大型语言模型(LLMs)已经改变了软件工程领域。最近,众多基于LLM的代理已被开发出来,以解决现实中的软件问题修复任务。尽管其最先进的性能,这些代理面临一个重大挑战:高质量问题描述不足。现实数据集常出现问题描述与其对应解决方案之间的不一致,引入噪声和模糊性,导致自动化代理误解并限制其解决问题的效果。我们提出了SWE-Fuse,一种问题描述感知的训练框架,融合了问题描述引导和无问题样本来训练SWE代理。它包含两个关键模块:(1)一个无问题驱动的轨迹学习模块,用于缓解可能误导的问题描述,同时使模型学习逐步调试过程;(2)一个熵感知的RLVR训练模块,通过熵驱动的剪枝自适应调整训练动态。它在高熵下应用松驰剪枝以鼓励探索,在低熵下应用严格剪枝以确保训练稳定性。我们在广泛研究的SWE-bench验证基准上评估SWE-Fuse,以证明其在解决现实软件问题上的有效性。具体而言,SWE-Fuse在解决率上分别优于最佳的8B和32B基线模型43.0%和60.2%。此外,将SWE-Fuse与测试时扩展(TTS)结合,可进一步提升性能,在TTS@8下,8B和32B模型的解决率分别达到49.8%和65.2%。

英文摘要:

Large language models (LLMs) have transformed the software engineering landscape. Recently, numerous LLM-based agents have been developed to address real-world software issue fixing tasks. Despite their state-of-the-art performance, Despite achieving state-of-the-art performance, these agents face a significant challenge: \textbf{Insufficient high-quality issue descriptions.} Real-world datasets often exhibit misalignments between issue descriptions and their corresponding solutions, introducing noise and ambiguity that mislead automated agents and limit their problem-solving effectiveness. We propose \textbf{\textit{SWE-Fuse}}, an issue-description-aware training framework that fuses issue-description-guided and issue-free samples for training SWE agents. It consists of two key modules: (1) An issue-free-driven trajectory learning module for mitigating potentially misleading issue descriptions while enabling the model to learn step-by-step debugging processes; and (2) An entropy-aware RLVR training module, which adaptively adjusts training dynamics through entropy-driven clipping. It applies relaxed clipping under high entropy to encourage exploration, and stricter clipping under low entropy to ensure training stability. We evaluate SWE-Fuse on the widely studied SWE-bench Verified benchmark shows to demonstrate its effectiveness in solving real-world software problems. Specifically, SWE-Fuse outperforms the best 8B and 32B baselines by 43.0\% and 60.2\% in solve rate, respectively. Furthermore, integrating SWE-Fuse with test-time scaling (TTS) enables further performance improvements, achieving solve rates of 49.8\% and 65.2\% under TTS@8 for the 8B and 32B models, respectively.

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。

↑