让AI自主优化记忆策略，大幅提升长时任务表现。
SelfMem: Self-Optimizing Memory for AI Agents
- AI自主探索并优化记忆存储与检索方式，无需人工设定规则。
- 在10万至100万词的对话中，性能比最强基线提升超40%。
- 适合需要长期记忆、复杂推理的智能体应用开发。
尽管当前AI代理支持越来越长的上下文窗口、工具调用和技能执行以完成长周期任务，但仍需有效的记忆系统来利用历史经验。现有记忆框架通常依赖固定的记忆存储、检索和摘要机制，这些机制在不同任务间僵化且常需手动调优。为此，我们提出SelfMem——一种自优化记忆框架。受自改进AI研究启发，遵循‘授人以渔而非授人以鱼’的原则，不强制模型遵循预定义记忆策略或格式，而是提供包含记忆工具和反馈信号的环境，使代理能够自主探索、评估并优化自身记忆策略。结果表明，SelfMem在BEAM基准上，跨10万至100万词规模的对话中持续优于检索、压缩及代理-记忆基线。相较于最强基线，其在10万、50万和100万词规模下的官方得分分别提升48.7%、40.8%和41.9%。进一步的问题类型分析显示其在多样化记忆需求下具有广泛鲁棒性，优化研究也证明模型引导的策略精炼可进一步提升性能。
原文摘要 · Abstract (English)
While current AI agents support increasingly long context windows, tool use, and skill execution for long-horizon tasks, they still require memory systems to effectively leverage historical experience. Existing memory frameworks typically rely on fixed storage, retrieval, and summarization mechanisms, which can be rigid across different tasks and often require manual tuning. To address this limitation, we propose SelfMem, a self-optimizing memory framework. Inspired by prior work on self-improving AI, we follow the principle of "teaching an agent to fish rather than giving it a fish." Instead of forcing the model to follow a predefined memory strategy or format, SelfMem provides an environment with memory tools and feedback signals that allow the agent to explore, evaluate, and refine its own memory strategy. Our results show that SelfMem consistently outperforms retrieval, compression, and agent-memory baselines on BEAM across conversation scales from 100K to 1M tokens. Compared with the strongest baseline, SelfMem improves the official score by 48.7%, 40.8%, and 41.9% at 100K, 500K, and 1M, respectively. Further question-type analysis shows broad robustness across diverse memory demands, and our optimization study shows that model-guided strategy refinement further improves performance.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。