让AI学会按用户需求选择记忆，避免信息堆积导致回答变差。
AdaMem: Learning What to Remember for Personalized Long-Horizon LLM Agents
- 根据用户反馈动态学习该记住什么，而非全量保存对话内容。
- 在基准测试中提升问答准确率最高达9.0%，内存减少9%。
- 适合需要长期个性化交互的AI助手场景，如客服、个人助理。
大型语言模型（LLM）代理的长期记忆系统通常试图「记住一切」，将对话内容均匀提取为记忆。但在实际应用中，推理成本和有限的上下文预算使这种做法不可行：除了将原始对话整合进记忆外，代理还需具备「写入控制」能力，仅保留每个用户真正关心的信息。否则，长期个性化交互将遭遇「记忆膨胀」问题——无关琐事挤占有效信息，持续降低问答准确率。我们提出，值得记忆的内容具有角色依赖性，因此设计了自适应记忆机制AdaMem，通过每周的问答反馈学习每位用户的记忆偏好。AdaMem维护结构化的角色专属记忆策略，并通过轻量级、分块式的自我反思与失败回滚机制进行优化。为评估该设定，我们构建了AdaMem-Bench基准，模拟多周交互及逐周问答反馈。在两种提取模型和两种反馈模式下，相比统一记忆基线Mem0，AdaMem最高提升问答准确率9.0%，同时内存体积减少9%。
原文摘要 · Abstract (English)
Long-term memory systems for Large Language Model (LLM) agents typically try to \emph{remember everything}, extracting memories uniformly to retain as many facts as possible. In production, however, inference cost and finite context budgets make this untenable: beyond consolidating raw dialogue into memory, an agent must exert \emph{write control}, efficiently keeping only the information each user actually cares about. Otherwise, long-horizon personalized interactions suffer \emph{memory bloat}, where irrelevant trivia crowds out useful information and steadily erodes question-answering (QA) accuracy. We argue that what is worth remembering is role-dependent, and propose \textbf{AdaMem} (Adaptive Memory), a method that \emph{learns what to remember} for each user from feedback. AdaMem maintains a structured, role-specific Memory Policy and refines it from weekly QA feedback through a lightweight, patch-style self-reflection step with failure rollback. To study this setting, we build \textbf{AdaMem-Bench}, a benchmark that simulates weeks of interaction with week-by-week QA. Across two extraction models and two feedback modes, AdaMem improves QA accuracy by up to \textbf{+9.0\%} over the uniform Mem0 baseline while shrinking memory volume by \textbf{9\%}.
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。