[论文] MedGraph: Maintaining Patient State with Graph Memory for Long-Term Me...

论文概要

研究领域: NLP 作者: Donghua Zhang, Zizheng Wang, Ze Dong, Yiding Feng, Siyu Qian 发布时间: 2026-11-05 arXiv: 2611.08433

中文摘要

基于 LLM 的医疗助手需要跨就诊跟踪患者，因此需要一种在患者临床状态变化时保持正确性的记忆。我们考察了五个记忆系统在纵向医疗对话上的错误，发现三个反复出现的原因：检索到的证据给出了过时的状态、包含了他人的健康事实、或缺少答案所依赖的安全事实。我们提出 MedGraph，一个针对这三种情况构建的图记忆。构建阶段从每次会话中提取带日期的事实，将每个事实归属到其描述的人，并将用药、诊断和测量记录链接成保留早期状态的链。检索阶段将患者的链作为单元排序，在问题所问及的时间点读取每条链，并添加原始对话文本；患者档案随每个问题列出所有已记录的安全事实。在 MedMemoryBench 上，MedGraph 在使用 GPT 和 Qwen 作为骨干模型时都取得了最高总体准确率，分别比五个基线中最强的高出 10.5 和 10.4 个百分点，同时在 GPT 设置下记忆构建和回答效率也更高。

原文摘要

LLM-based medical assistants that follow a patient across visits need a memory that keeps the patient's clinical state correct as it changes. We examine the errors of five memory systems on longitudinal medical dialogue and find three recurring causes: the retrieved evidence gives an outdated state, includes another person's health facts, or lacks a safety fact the answer depends on. We present MedGraph, a graph memory built for these three cases. Construction extracts dated facts from each session, attributes each fact to the person it describes, and links medication, diagnosis, and measurement records into chains that keep earlier states. Retrieval ranks the patient's chains as units, reads each chain at the time the question asks about, and adds the original dialogue text; a patient profile lists every recorded safety fact with every question. On MedMemoryBench, MedGraph has the highest overall accuracy with both a GPT and a Qwen backbone, 10.5 and 10.4 points above the strongest of five baselines, with efficient memory construction and answering in the GPT setting.

---

*自动采集于 2026-11-08*

#论文 #arXiv #NLP #小凯