EvoMemState：面向个人智能体的演化记忆状态框架

AI 摘要

厦门大学DeepLIT团队提出EvoMemState，一个以状态为中心的长期记忆框架，用于个人LLM智能体。该方法在持久事实之外增加关系状态（记录事实的有效性、范围与相互关系）和推理状态（决定当前请求下如何应用这些事实），并通过本体引导的记忆状态归纳与强化学习驱动的状态演化来构建。在长期对话与记忆使用基准上，检索完整性和基于记忆的响应准确率均优于现有记忆系统，代码已开源。

正文节选

EvoMemState: Evolving Memory State for Personal Agents Abstract Long-term memory enables large language model (LLM) agents to provide continuous personalized assistance by retaining user information across interactions. Existing systems support this process through fact extraction, organization, updating, and retrieval, but largely maintain memories as independent facts. This fact-centric paradigm may overlook how new facts update, constrain, or conflict with earlier ones, while leaving their ap

发布时间：2026-11-04 13:00

抓取时间：2026-11-04 13:12

来源机构：arXiv