arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2610.21170cs.AI
收藏

MemScope:通过本地化代码修复演进智能体记忆架构

MemScope: Evolving Agent Memory Architectures through Localized Code Repair

State Key Laboratory of AI Safety(人工智能安全国家重点实验室);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);University of Chinese Academy of Sciences(中国科学院大学);Meituan(美团)

Miao Su, Yucan Guo, Zixuan Li, Long Bai, Zhongni Hou, Xiaolong Jin, Jiafeng Guo

AI总结:

MemScope是基于证据的智能体记忆架构本地化演进框架,通过Patch Agent联合检查执行轨迹、记忆状态与代码实现精准修复,在GAIA、xBench等数据集上显著提升了智能体任务性能。

摘要:

智能体记忆架构定义了大语言模型(LLM)智能体如何在不同任务间存储和复用经验,但这类架构通常在部署后便固定不变,限制了其在各类任务类型中的适应性。近期的自演进方法转而利用执行反馈来调整记忆架构,这类反馈通常表现为评估分数或压缩摘要。然而,这种粗粒度的反馈不足以在当前智能体系统中实现精准的故障定位,因为长周期任务执行涉及多个相互作用的组件。若缺乏精准的故障定位,现有方法便缺少明确的代码级修改目标,导致更新范围约束薄弱,往往会对原本可行的记忆架构进行大规模重写。我们提出MemScope,一个基于证据的现有记忆架构本地化演进框架。其使用工具的Patch Agent会联合检查原始执行轨迹、运行时记忆状态和当前架构代码,以识别修复目标,并在提供者身份和记忆兼容性约束下应用有界的原位编辑。在GAIA数据集上,结合轨迹与记忆检查的变体达到了86.67%的准确率,而仅使用轨迹反馈的变体准确率为70.00%,仅使用摘要反馈的变体准确率为63.33%。在xBench任务上,经演进的记忆架构在使用DeepSeek V3.2时,正确答案数量从49提升至55;在使用Kimi K2.7 Code时,正确答案数量从50提升至53。在不相交的ALFWorld未见过的任务上,成功率从50.75%提升至71.64%。在已归档的架构快照转换中,代码变更比例的中位数为4.4%。这些结果表明,细粒度的执行证据可作为本地化、原位记忆架构演进的基础。

英文摘要:

Agent memory architectures define how LLM agents store and reuse experience across tasks, but they are typically fixed after deployment, limiting their adaptability across task types. Recent self-evolving approaches instead adapt these memory architectures using execution feedback, typically represented as evaluation scores or compressed summaries. However, such coarse-grained feedback is insufficient for precise failure localization in current agent systems, where long-horizon task execution involves multiple interacting components. Without a precise failure location, existing approaches lack a clear code-level target for modification, leaving the update scope weakly constrained and often leading to large-scale rewrites of an otherwise capable memory architecture. We propose MemScope, a framework for evidence-grounded, localized evolution of an existing memory architecture. Its tool-using Patch Agent jointly inspects raw execution traces, runtime memory state, and current architecture code to identify repair targets and apply bounded, in-place edits under provider-identity and memory-compatibility constraints. On GAIA, the trace-and-memory inspection variant achieves 86.67% accuracy, compared with 70.00% for trace-only and 63.33% for summary-only feedback. Under xBench tasks, the evolved memory architecture improves over its original version from 49 to 55 correct answers with DeepSeek V3.2 and from 50 to 53 with Kimi K2.7 Code. On disjoint ALFWorld unseen tasks, success rises from 50.75% to 71.64%. Across archived architecture-snapshot transitions, the median code-change ratio is 4.4%. These results support fine-grained execution evidence as a basis for localized, in-place memory-architecture evolution.

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。

↑