SWE-Protégé: 通过选择性协作专家解锁小型语言模型作为软件工程代理
SWE-Protégé: Learning to Selectively Collaborate With an Expert Unlocks Small Language Models as Software Engineering Agents
AI总结：
SWE-Protégé通过选择性协作专家，使小型语言模型在软件工程任务中取得显著提升，实现42.4%的Pass@1成绩。
摘要：
小型语言模型（SLMs）在成本、延迟和适应性方面具有显著优势，但迄今为止在长期软件工程任务（如SWE-bench）上仍落后于大模型，因为它们在动作循环和低解决率方面存在普遍问题。我们介绍了SWE-Protégé，一种在训练后框架，将软件修复重新定义为专家-学徒协作问题。在SWE-Protégé中，SLM仍然是唯一的决策者，同时学习选择性地寻求强大专家模型的指导，识别停滞状态，并遵循专家反馈。我们的方法结合了对专家增强轨迹的监督微调和代理强化学习，后者明确 discourages 退化循环和无生产力的专家协作。我们轻量级地在Qwen2.5-Coder-7B-Instruct上进行后训练，以在SWE-bench Verified上达到42.4%的Pass@1，比先前的SLM最佳状态提高了25.4%，同时使用专家帮助稀疏（每个任务约4次调用和总token的11%）
英文摘要：
Small language models (SLMs) offer compelling advantages in cost, latency, and adaptability, but have so far lagged behind larger models on long-horizon software engineering tasks such as SWE-bench, where they suffer from pervasive action looping and low resolution rates. We introduce SWE-Protégé, a post-training framework that reframes software repair as an expert-protégé collaboration problem. In SWE-Protégé, an SLM remains the sole decision-maker while learning to selectively seek guidance from a strong expert model, recognize stalled states, and follow through on expert feedback. Our approach combines supervised fine-tuning on expert-augmented trajectories with agentic reinforcement learning that explicitly discourages degenerative looping and unproductive expert collaboration. We lightly post-train Qwen2.5-Coder-7B-Instruct to achieve 42.4% Pass@1 on SWE-bench Verified, a +25.4% improvement over the prior SLM state of the art, while using expert assistance sparsely (~4 calls per task and 11% of total tokens).
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。
