arXiv cs.CL· Dongfang Li, Zhuoen Chen, Bonian Jia, Zixuan Liu, Zhengxuan Lu, Can Deng, Yukun Shi, Yuhang Meng, Xin Zhang, Jiahe Huang, Fuhao Li, Junmai Wang, Shunyao Wen, Jingjun Sun, Jiaxu Tian, Jiangling Zhang, Zihan Wang, Liangyou Li, Yitong Yao, Shuangyong Song, Beilun Wang, Pengxu Wei, Hanlin Xu, Chang Liu, Yongxiang Li, Baotian Hu, Xiangyang Ji, Min Zhang·· 13 小时前AI 评分36
以记忆为中心的持续学习（MCCL）：让 AI 智能体从被动检索走向持续学习
Memory-Centric Continual Learning for AI Agents
AI 导读
研究者提出 Memory-Centric Continual Learning（MCCL），主张记忆应以是否改善未来行动而非检索准确率来评判。MCCL 包含三轴记忆分类法（存储载体、资产形式、管理策略）与内外双环生命周期，并配套干扰下保持、变化适应、任务迁移的评测协议及记忆扩展框架。
正文 · 原文
Authors:Dongfang Li, Zhuoen Chen, Bonian Jia, Zixuan Liu, Zhengxuan Lu, Can Deng, Yukun Shi, Yuhang Meng, Xin Zhang, Jiahe Huang, Fuhao Li, Junmai Wang, Shunyao Wen, Jingjun Sun, Jiaxu Tian, Jiangling Zhang, Zihan Wang, Liangyou Li, Yitong Yao, Shuangyong Song, Beilun Wang, Pengxu Wei, Hanlin Xu, Chang Liu, Yongxiang Li, Baotian Hu, Xiangyang Ji, Min Zhang
Abstract:Autonomous agents powered by foundation models can reason, plan, and use tools, but their fixed weights and context limits hinder sustained learning from experience over long-term deployment. While external memory systems attempt to bridge this gap, they mostly act as passive retrieval caches, evaluated primarily by whether they can recall stored facts or past interactions. Whether memory can drive genuine continual learning remains an open question. In this paper, we introduce Memory-Centric Continual Learning (MCCL), arguing that memory should be judged by whether it improves future actions, not merely what it retrieves. MCCL has three main parts. First, a three-axis taxonomy organizes memory by where state is stored (substrates), what form it takes (assets), and how it is managed (policies). This defines a two-loop lifecycle that updates memory state in an inner loop while refining management policies in an outer loop based on task feedback. Second, an evaluation protocol measures retention under interference, adaptation to change, and task transfer to isolate genuine behavioral gains from simple context expansion. Third, a memory scaling framework models how agent performance scales with experience, storage capacity, and compute budgets across non-parametric and parametric memory. Together, these components reframe agent memory from passive lookup to continuous learning, offering a principled route toward self-improving autonomous agents.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.29357 [cs.CL] |
| (or arXiv:2610.29357v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29357 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dongfang Li [view email]
[v1]
Wed, 21 Oct 2026 07:49:42 UTC (4,176 KB)
来源：arXiv cs.CL · arxiv.org