arXiv cs.CL· Zehao Lin, Erenkai Zhu, Helan Hu, Xixuan Hao, Chunyu Li, Kai Chen·· 13 小时前AI 评分34
语言智能体的记忆价值：一个决策论框架
The Value of Memory in Language Agents
AI 导读
研究者提出一个决策论框架，用“避免任务损失”为单位衡量语言智能体记忆的原始价值与保留价值，并刻画风险-成本前沿。该框架通过成本感知模拟和五个最小反例，揭示信息时序、检索访问、组件交互、反馈与需求不确定性各自的作用，其中一个与答案互信息为零的指针仍能降低精确恢复成本。在 LoCoMo、双项目更新任务和纵向案例上的诊断显示，记忆在保留、使用与测量三个环节均存在失效。
正文 · 原文
Abstract:Memory enables language agents to carry useful distinctions from past experience into future decisions. Its value depends on which distinctions survive, whether a bounded agent can use them, and what can be recovered or revised as demands evolve. We develop a decision-theoretic framework that follows this value from raw history through persistent state, operational use, and costly recovery and revision. Original and retained value are expressed in units of avoided task loss; risk-cost frontiers describe attainable recovery and preservation across uses. Unknown demand yields value profiles, while interactive feedback changes the distribution of future work. A cost-aware simulation condition and five minimal counterexamples expose the roles of information timing, retrieval access, component interaction, feedback and demand uncertainty. In particular, a pointer with zero mutual information about the answer can still reduce the cost of exact recovery. Diagnostics on LoCoMo, a two-project update task and longitudinal cases exhibit failures of preservation, use and measurement. At fixed retrieved evidence, a saved ledger changes subsequent answers; at fixed agent outputs, an evaluator revision changes scores. The framework connects these distinctions to operational risk bounds and testable principles for allocating memory, preserving recovery options and revising evidence dependencies.
| Comments: | 19 pages, 3 figures, 5 tables |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.31303 [cs.AI] |
| (or arXiv:2610.31303v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31303 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zehao Lin [view email]
[v1]
Thu, 22 Oct 2026 14:01:31 UTC (158 KB)
来源：arXiv cs.CL · arxiv.org