arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2607.01935cs.AI
收藏

A-TMA: 解耦长期智能体记忆中的状态感知记忆故障

A-TMA: Decoupling State-Aware Memory Failures in Long-Term Agent Memory

National University of Singapore(新加坡国立大学)

Zitong Shi, Yixuan Tang, Anthony Kum Hoe Tung

AI总结:

针对长期记忆中状态不一致导致的幽灵记忆问题,提出A-TMA方法,通过状态感知覆盖层在存储、检索和回答三个层面解耦故障,在冲突密集基准上显著提升时间F1。

摘要:

长期记忆使LLM智能体能够作为持久助手,但用户事实会发生变化。一个有用的记忆系统必须知道当前什么是真的、过去什么是真的,以及什么发生了变化。我们研究\emph{幽灵记忆},一种状态协调故障,其中旧事实、当前事实和过渡事实共存于记忆库中,在检索时仍然混合,并误导回答模型。我们认为记忆系统应从三个层面理解和优化:库维护、检索和回答时间解析。我们提出ATMA,一种针对现有记忆系统的状态感知覆盖层。ATMA在库中保留被取代的记录和过渡记录,为查询请求的状态视图构建证据包,并向QA暴露当前、历史和过渡标签。我们进一步呼吁对库、检索和回答层面的故障进行解耦评估,因为最终QA准确性可能隐藏幽灵记忆发生的位置。为了使这种故障可测量,我们构建了LTP(LoCoMo Temporal Plus),一个针对幽灵记忆的冲突密集型基准,并在LoCoMo上评估长期对话泛化能力。在LTP上,Graphiti+ATMA相比Graphiti将冲突准确性绝对提高了0.240。在LoCoMo上,Graphiti+ATMA将时间F1从0.0295提升至0.1705。这些增益依赖于宿主,但表明显式状态角色可以减少被最终QA准确性隐藏的记忆故障。

英文摘要:

Long term memory lets LLM agents act as persistent assistants, but user facts change. A useful memory system must know what is true now, what used to be true, and what changed. We study \emph{ghost memory}, a state coordination failure in which old, current, and transition facts coexist in the memory bank, remain mixed during retrieval, and mislead the answer model. We argue that memory systems should be understood and optimized from three levels: bank maintenance, retrieval, and answer time resolution. We propose ATMA, a state aware overlay for existing memory systems. ATMA keeps superseded and transition records in the bank, builds evidence packets for the query's requested state view, and exposes current, historical, and transition labels to QA. We further call for decoupled evaluation of bank, retrieval, and answer level failures, since final QA accuracy can hide where ghost memory occurs. To make this failure measurable, we build LTP (LoCoMo Temporal Plus), a conflict heavy benchmark for ghost memory, and evaluate on LoCoMo for long conversation generalization. On LTP, Graphiti+ATMA improves conflict accuracy by 0.240 absolute over Graphiti. On LoCoMo, Graphiti+ATMA raises temporal F1 from 0.0295 to 0.1705. The gains are host dependent, but they indicate that explicit state roles can reduce memory failures hidden by final QA accuracy.

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。

↑