[论文] UniMem: Complementary Episodic-to-Parametric Memory for Boundary-Agnos…

## 论文概要 **研究领域**: NLP **作者**: Siyu Xia, Chenheng Zhang,...

论文概要

研究领域: NLP 作者: Siyu Xia, Chenheng Zhang, Yanting Wu, Haoxuan Li, Jiajun Chai, Xiaohan Wang, Guojun Yin, Wei Lin, Zhouchen Lin, Haifeng Zhang, Jun Wang 发布时间: 2026-07-28 arXiv: 2607.26017

中文摘要

记忆对于LLM智能体积累任务经验和复用任务特定执行策略至关重要。然而,在边界不可知且不断演化的任务流上的实际部署暴露了一个根本性的稳定-可塑性困境。基于外部检索的记忆可以快速吸收新证据,但往往无法内化重复的执行模式,并带来推理时的检索开销。参数化记忆一旦学会就能实现稳定高效的执行,但通常依赖于显式任务边界和固定参数预算。受人类大脑通过互补的情景存储和渐进巩固来平衡可塑性和稳定性的启发,我们提出了UniMem,一个用于自主记忆管理的自路由框架。UniMem使用可学习的路由token作为记忆控制器,实现互补记忆路径的自适应协调:新颖或稀疏任务保留在情景缓冲区中用于检索增强执行,而重复且可靠的模式被巩固到可扩展的参数化记忆中。通过使用路由token和参数化记忆块将任务识别与任务执行解耦,UniMem在部署期间无需任务标签即可按需扩展记忆,且不会出现不受控的参数增长。在长程流式任务序列上的实验表明,UniMem始终优于基线,同时保持执行保真度,在三个骨干模型上平均获得4.0 EM点的提升。

原文摘要

Memory is essential for LLM agents to accumulate task experience and reuse task-specific execution strategies. However, real-world deployment over boundary-agnostic and evolving task streams exposes a fundamental stability-plasticity dilemma. External retrieval-based memory can rapidly absorb new evidence, but it often fails to internalize recurring execution patterns and incurs inference-time retrieval overhead. Parametric memory enables stable and efficient execution once learned, but typically relies on explicit task boundaries and fixed parameter budgets. Inspired by the human brain, which balances plasticity and stability through complementary episodic storage and gradual consolidation, we propose UniMem, a self-routing framework for autonomous memory management. UniMem uses learnable…

自动采集于 2026-07-30

#论文 #arXiv #NLP #小凯

发表回复

人生梦想 - 关注前沿的计算机技术 acejoy.com 🐾 步子哥の博客 🐾 背多分论坛 🐾 借一步网 🐾 智柴网 沪ICP备2024052574号-1