发表机构
Institute of Automation, Chinese Academy of Sciences; School of Artificial Intelligence, University of Chinese Academy of Sciences; State Key Lab of General Artificial Intelligence, School of Intelligence Science and Technology, Peking University; Institute for Artificial Intelligence, Peking University; Meituan; AI Centre, Department of Computer Science, University College London（中国科学院自动化研究所; 中国科学院大学人工智能学院; 北京大学智能科学与技术学院通用人工智能国家重点实验室; 北京大学人工智能研究院; 美团; 伦敦大学学院计算机科学系人工智能中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对大语言模型在无边界任务流部署的稳定性 - 可塑性困境，提出UniMem自路由框架，用可学习路由令牌协调互补记忆路径，能按需扩展记忆，实验证明其在长周期流任务序列中优于基线并保持执行保真度。
AI 中文摘要
记忆对于大语言模型智能体积累任务经验和复用特定任务执行策略至关重要。然而，在无边界且不断演变的任务流上进行实际部署时，会出现基本的稳定性 - 可塑性困境。基于外部检索的记忆能快速吸收新证据，但常无法内化重复执行模式且产生推理时的检索开销。参数记忆一旦学习完成就能实现稳定高效执行，但通常依赖明确的任务边界和固定参数预算。受人类大脑通过互补情景存储和逐步巩固来平衡可塑性和稳定性的启发，我们提出了UniMem，一种用于自主记忆管理的自路由框架。UniMem使用可学习的路由令牌作为记忆控制器，在互补记忆路径间实现自适应协调。新的或稀疏的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则被巩固到可扩展的参数记忆中。通过用路由令牌和参数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或不受控制的参数增长就能按需扩展记忆。在长周期流任务序列上的实验表明，UniMem在保持执行保真度的同时始终优于基线，在三个骨干模型上平均获得4.0 EM点的增益。
英文摘要
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 routing tokens as memory controllers, enabling adaptive coordination between complementary memory pathways: novel or sparse tasks are retained in an episodic buffer for retrieval-augmented execution, while recurring and reliable patterns are consolidated into expandable parametric memory. By decoupling task identification from task execution with routing tokens and parametric memory blocks, UniMem expands memory on demand without task labels during deployment or uncontrolled parameter growth. Experiments on long-horizon streaming task sequences show that UniMem consistently outperforms baselines while maintaining execution fidelity, achieving an average gain of 4.0 EM points across three backbone models.
