---
id: 20260924-T0-11
title: "MoM：让LLM agent记住\"当前有效\"的记忆，而非堆积历史记录"
title_en: "MoM: Memory Architecture for LLM Agents That Tracks What Currently Holds"
url: https://ai.daily.yangsir.net/daily/20260924-T0-11
issue_date: 2026-09-24
publish_date: 2026-09-23T04:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2609.25054
---

# MoM：让LLM agent记住"当前有效"的记忆，而非堆积历史记录

研究者提出MoM（Memory of Memory）方法，解决长周期LLM agent的记忆问题。现有方案通常存储所有交互记录，在查询时重构当前状态；MoM认为关键不是"曾记录了什么"，而是"当前什么仍然成立"。该方法为agent记忆管理提供了新的思路，论文发布于arXiv。

## English Version

**MoM: Memory Architecture for LLM Agents That Tracks What Currently Holds**

Researchers proposed MoM (Memory of Memory), addressing the memory problem for long-horizon LLM agents. Most existing designs store every interaction and reconstruct the present at query time; MoM argues the key question is not what was once recorded but what currently holds. The paper offers a new approach to agent memory management and is available on arXiv.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2609.25054)

**详情页**：https://ai.daily.yangsir.net/daily/20260924-T0-11

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*