arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.01252cs.CLcs.AI
收藏

将知识与关怀联系起来:知识图谱增强的医疗随访问题生成

Linking Knowledge to Care: Knowledge Graph-Augmented Medical Follow-Up Question Generation

Carnegie Mellon university(卡内基梅隆大学);Amazon Health AI(亚马逊健康AI)

Liwen Sun, Xiang Yu, Ming Tan, Zhuohao Chen, Anqi Cheng, Ashutosh Joshi, Chenyan Xiong

AI总结:

本文提出KG-Followup方法,通过知识图谱增强LLM,提升医疗随访问题生成的准确性和效率。

摘要:

临床诊断耗时长,需要患者与医疗专业人员之间进行密集的互动。尽管大型语言模型(LLMs)可以减轻预诊断工作的负担,但其有限的领域知识阻碍了有效的医疗问题生成。我们引入了一个知识图谱增强的LLM,结合主动上下文学习,生成相关且重要的随访问题,称为KG-Followup,作为预诊断评估的关键模块。结构化的医疗领域知识图谱充当无缝的补丁,为LLM提供专业领域的专业知识,使其能够推理。实验表明,KG-Followup在相关基准测试中比最先进的方法在召回率上高出5%至8%。

英文摘要:

Clinical diagnosis is time-consuming, requiring intensive interactions between patients and medical professionals. While large language models (LLMs) could ease the pre-diagnostic workload, their limited domain knowledge hinders effective medical question generation. We introduce a Knowledge Graph-augmented LLM with active in-context learning to generate relevant and important follow-up questions, KG-Followup, serving as a critical module for the pre-diagnostic assessment. The structured medical domain knowledge graph serves as a seamless patch-up to provide professional domain expertise upon which the LLM can reason. Experiments demonstrate that KG-Followup outperforms state-of-the-art methods by 5% - 8% on relevant benchmarks in recall.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑