UMEM: 通用记忆的统一记忆提取与管理框架
UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory
AI总结：
UMEM通过联合优化大语言模型，实现记忆的提取与管理，提升记忆的泛化能力，实验显示在多轮交互任务中表现优于基线模型。
摘要：
自我进化记忆作为大语言模型（LLMs）基于代理的可训练参数，其中提取（从经验中提炼洞察）和管理（更新记忆库）必须紧密协调。现有方法主要优化记忆管理，而将记忆提取视为静态过程，导致泛化能力差，其中代理积累实例特定的噪声而非稳健的记忆。为了解决这个问题，我们提出了统一记忆提取与管理（UMEM），一个自我进化代理框架，联合优化一个大语言模型以同时提取和管理记忆。为了缓解对特定实例的过拟合，我们引入了语义邻域建模，并通过GRPO优化模型，采用邻域级边际效用奖励。这种方法通过在语义相关查询簇中评估记忆效用，确保记忆的泛化能力。在五个基准上的广泛实验表明，UMEM显著优于高度竞争的基线，实现了在多轮交互任务中高达10.67%的提升。此外，UMEM在连续进化过程中保持单调增长曲线。代码和模型将公开发布。
英文摘要：
Self-evolving memory serves as the trainable parameters for Large Language Models (LLMs)-based agents, where extraction (distilling insights from experience) and management (updating the memory bank) must be tightly coordinated. Existing methods predominately optimize memory management while treating memory extraction as a static process, resulting in poor generalization, where agents accumulate instance-specific noise rather than robust memories. To address this, we propose Unified Memory Extraction and Management (UMEM), a self-evolving agent framework that jointly optimizes a Large Language Model to simultaneous extract and manage memories. To mitigate overfitting to specific instances, we introduce Semantic Neighborhood Modeling and optimize the model with a neighborhood-level marginal utility reward via GRPO. This approach ensures memory generalizability by evaluating memory utility across clusters of semantically related queries. Extensive experiments across five benchmarks demonstrate that UMEM significantly outperforms highly competitive baselines, achieving up to a 10.67% improvement in multi-turn interactive tasks. Futhermore, UMEM maintains a monotonic growth curve during continuous evolution. Codes and models will be publicly released.
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。
