MemLens: A Value-Aware Memory Management System with Interactive Analytics for LLM-based Agents

ID: f68c6225
原文链接: https://arxiv.org/abs/2607.25992
PDF: https://arxiv.org/pdf/2607.25992v1
作者: Shuyue Wei, Chang Liu, Zimu Zhou, Yongxin Tong, Lizhen Cui
日期: 2026-07-28
更新: 2026-07-28
分类: agents
来源类型: paper
标签: agent-memory, memory-management, value-aware, interactive-analytics, shapley
质量评分: 4/5
抓取时间: 2026-07-30T12:30:29Z
arXiv 分类: cs.DB, cs.AI
Comment: N/A

中文导读

MemLens 提出 value-aware 的 agent 记忆管理系统，将记忆条目作为一等公民数据对象，而非均匀存储异构交互记录提供端到端交互分析面板，覆盖完整记忆生命周期：Shapley 式记忆价值评估价值感知存储记忆辅助响应通过 study-copilot 应用展示了记忆价值检查层次结构可视化及多策略对比（响应质量/检索延迟/token 消耗）

为什么值得关注

用 Shapley 估记忆价值：MemLens 让 agent 记忆从粗粒度存储升级为可检查可比较的价值感知系统

摘要强调现有 LLM memory 对异构交互记录粗粒度等待，MemLens 用 Shapley 式价值评估 + 交互分析面板支持策略对比（质量/延迟/token）。

关键信息

论文标题：MemLens: A Value-Aware Memory Management System with Interactive Analytics for LLM-based Agents
作者：Shuyue Wei, Chang Liu, Zimu Zhou, Yongxin Tong, Lizhen Cui
arXiv：https://arxiv.org/abs/2607.25992
发布时间：2026-07-28
arXiv 分类：cs.DB, cs.AI
关联标签：agent-memory, memory-management, value-aware, interactive-analytics, shapley

English Abstract

Recently, memory management has become a key infrastructure for LLM-based agents, as it directly affects long-horizon reasoning, personalized responses, and knowledge reuse. However, existing LLM memory systems typically adopt a coarse-grained (utility-agnostic) manner that treats heterogeneous user-LLM interaction records uniformly, leading to redundant and low-impact records persisting in the memory repository. To address this challenge, we present MemLens, a value-aware memory management system that takes memory records as first-class data objects. MemLens provides an end-to-end interactive analytics dashboard that exposes the complete memory lifecycle, including Shapley-style memory evaluation, value-aware storage, and memory-assisted response. Through a study-copilot application, the system enables users to inspect memory values, visualize hierarchical memory structures, and compare various memory management strategies in terms of response quality, retrieval latency, and token consumption. Therefore, our MemLens can serve as an efficient, interpretable, and personalized long-term memory management system for LLM-based agents.

English Summary

MemLens is a value-aware memory management system for LLM-based agents that treats memory records as first-class data objects rather than uniformly storing heterogeneous interaction records. It provides an end-to-end interactive analytics dashboard exposing the complete memory lifecycle: Shapley-style memory evaluation, value-aware storage, and memory-assisted response generation. Through a study-copilot application, users can inspect memory values, visualize hierarchical memory structures, and compare strategies for response quality, retrieval latency, and token consumption.

Obsidian Notes

内容由 opencli arxiv paper 拉取 arXiv 元数据与摘要生成。
中文导读与价值判断均锚定在条目已有摘要、论文摘要、作者、日期与分类信息上；未补充论文摘要之外的实验细节。
Grounded by abs-page / opencli metadata after opencli-first fetch.