paperbot · PL 论文追踪

全部论文 待读 收藏

P2 Exp-SWE-Agent: A Memory-Enhanced Coding Agent for Sequential Issue Resolution

Exp-SWE-Agent:面向顺序问题解决的记忆增强编码智能体
FSE Proceedings of the 34th ACM International Conference on the Foundations of Software Engineering2026匹配度 3/10 · 论文研究的是增强 Coding Agent 的持久记忆模块以提升问题解决率,属于 Agent 自身能力改进,而非将静态分析应用于 Agentic 程序或用 LLM/Agent 改进静态分析与形式化验证;且用户明确不感兴趣于以提升 LLM/Agent 能力为目标的研究。引用 0 inbox
Yixin Li, Zhengran Zeng, Rui Xie, Wei Ye, Shikun Zhang
crossref: 2026-09-12dblp: 2026-09-12

AI 速览

一句话:通过持久记忆与检索-行动-蒸馏循环,提升仓库级问题解决的通过率与效率。

本文面向长周期、仓库级软件工程工作流中智能体难以复用历史经验的问题,提出 Exp-SWE-Agent。该方法在 Mini-SWE-Agent 框架上增加轻量级持久记忆模块,通过持续的“检索-行动-蒸馏”循环,从过往问题解决轨迹中抽取结构化的事实性与经验性知识,并在后续任务中作为策略提示使用。作者在 SWE-Bench Pro 的顺序 8×25 子集上评估,问题解决率(Pass%)从 42.7% 提升至 47.5%,同时持续减少交互步数。结果表明,结构化持久记忆能有效提升智能体在长周期仓库级软件工程任务中的准确性与效率。

问题
长周期、仓库级软件工程任务中,编码智能体难以从过往问题解决轨迹中积累并复用经验,导致后续任务准确率与效率受限。
方法
在 Mini-SWE-Agent 上引入轻量级持久记忆模块,通过检索-行动-蒸馏循环抽取结构化事实与经验知识,并将其作为后续任务的策略提示。
评估
在 SWE-Bench Pro 的顺序 8×25 子集上与基线对比,问题解决率从 42.7% 提升到 47.5%,同时交互步数持续减少。
原文摘要(Abstract)

We present Exp-SWE-Agent, our submission to the FSE-AIWare 2026 MemComp challenge. Exp-SWE-Agent augments the Mini-SWE-Agent framework with a lightweight persistent memory module. Through a continuous retrieve-act-distill cycle, the agent extracts structured factual and experiential knowledge from past issue-resolution trajectories and uses them as strategic hints for subsequent tasks. Evaluated on a sequential 8 × 25 subset of SWE-Bench Pro, Exp-SWE-Agent increases the issue resolution rate (Pass%) from 42.7% to 47.5% while consistently reducing the number of interaction steps. These results demonstrate that structured persistent memory effectively enhances both the accuracy and efficiency of agents in long-horizon, repository-level software engineering workflows.

链接与引用

DOI 原文 · PDF(开放获取) · DBLP

BibTeX
@inproceedings{LiZXYZ26,
  title = {Exp-SWE-Agent: A Memory-Enhanced Coding Agent for Sequential Issue Resolution},
  author = {Yixin Li and Zhengran Zeng and Rui Xie and Wei Ye and Shikun Zhang},
  booktitle = {FSE},
  year = {2026},
  doi = {10.1145/3803437.3807663}
}