让大模型长任务决策更准，通过智能筛选记忆内容。
MemArbiter: Decision-Time Memory Arbitration for Long-Horizon LLM Agents
- 按功能分类记忆项，动态调节关键信息的优先级。
- 在ALFWorld上达92.5%成功率，比最强基线高25.4个百分点。
- 适合需要长期规划与记忆管理的任务场景。
大型语言模型（LLM）代理需保留并利用跨步骤信息以在长周期任务中保持行为连贯性。现有方法虽提升了记忆可访问性，但行动相关的信息仍可能因组织不当、优先级错配或呈现方式不佳而无法有效指导当前决策，我们称此为“记忆-行动差距”。本文提出MemArbiter，一种面向功能的动态记忆仲裁框架，解决由记忆管理引发的该差距问题。MemArbiter将交互历史分解为原子项，分入五类功能记忆库，并结合库级需求、项级相关性、焦点-背景表征及时间呈现门控机制，动态调控记忆显著性。在ALFWorld基准上，使用开源动作生成模型，在500和750令牌的统一每步记忆预算下，成功率达82.8%和92.5%，分别优于最强基线20.9和25.4个百分点。同时提升失败后恢复能力，减少错误动作重复与状态-动作循环。结果表明，功能感知的记忆仲裁能更有效地使可用信息指导行动。
原文摘要 · Abstract (English)
Large language model (LLM) agents must retain and use cross-step information to act coherently in long-horizon tasks. Existing methods improve memory accessibility, yet action-relevant information may still fail to guide the current decision because it is poorly formed, organized, prioritized, or presented. We call this post-access failure the Memory-Action Gap. We propose MemArbiter, a function-aware memory arbitration framework that addresses the memory-management-induced component of this gap. MemArbiter decomposes interaction histories into atomic items, organizes them into five functional Memory Banks, and combines bank-level demand, item-level relevance, focal-ambient representations, and a temporal presentation gate to dynamically control memory salience. We evaluate MemArbiter on ALFWorld against Flat Retrieval and Flat Recency under unified per-step memory budgets. With an open-weight action-generation model, MemArbiter achieves success rates of 82.8% and 92.5% under 500- and 750-token budgets, outperforming the strongest baseline by 20.9 and 25.4 percentage points, respectively. It also improves post-failure recovery and reduces failed-action repetition and state-action recurrence. These results show that function-aware memory arbitration enables accessible information to guide actions more effectively.
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。