arXiv cs.CL· Zhaoqian Xue, Anay Somani, Siddhartha Singh, Junting Mu, Zhongyin Wang, Haozheng Liu, Ayesha Khatun, Yeaeun Kwon, Renkai Ma, Siyuan Ma, Zhe He, Danxia Yu, Lingyao Li·· 2 天前AI 评分38
AskGLP1：从在线患者叙述构建 GLP-1 体验知识图谱的 LLM 智能体
AskGLP1: A Knowledge Graph LLM Agent for Exploring GLP-1 Experience from Online Patient Narratives
AI 导读
研究团队推出 AskGLP1，一个将在线患者叙述转化为知识图谱并回答用药体验问题的 LLM 智能体。该系统基于 MedDRA 术语，从近 850 万条 Reddit 帖子中筛选并提炼 163,135 条，结合 2,506 条 WebMD 评论，构建了包含 385,790 条药物副作用与用药目的关系的知识图谱。
正文 · 原文
Authors:Zhaoqian Xue, Anay Somani, Siddhartha Singh, Junting Mu, Zhongyin Wang, Haozheng Liu, Ayesha Khatun, Yeaeun Kwon, Renkai Ma, Siyuan Ma, Zhe He, Danxia Yu, Lingyao Li
Abstract:GLP-1 medications are increasingly used for weight management, and many people share their experiences online. These narratives show why people take the drugs and what they experience, but they remain unstructured and disconnected from formal pharmacovigilance. We therefore present AskGLP1, a system that transforms online patient narratives into a knowledge graph (KG) and answers questions about medication experiences. Through an LLM-assisted pipeline grounded in the MedDRA terminology, AskGLP1 screens nearly 8.5 million Reddit posts for medication experiences and distills 163,135 of them, with 2,506 WebMD reviews, into a KG of 385,790 drug-side-effect and drug-purpose relations. We find that GLP-1 experiences are source- and drug-specific. Online narratives share the most common gastrointestinal symptoms with FAERS, but they foreground everyday burdens such as fatigue and food aversion, whereas FAERS reports more often include injection-site reactions and selected serious diagnoses such as pancreatitis. We further find that posts report less nausea and vomiting for tirzepatide than for semaglutide, more fatigue for Wegovy than Ozempic and for Zepbound than Mounjaro. Evaluations show that the agent answers with verifiable evidence and declines requests for medical advice. Our findings reveal gaps between what FAERS captures and what people report online, and they call for patient-centered approaches to pharmacovigilance. Code, prompts, and evaluation questions are available at this https URL.
| Comments: | 16 pages, 4 figures, 11 tables |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI) |
| Cite as: | arXiv:2610.34987 [cs.CL] |
| (or arXiv:2610.34987v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.34987 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhaoqian Xue [view email]
[v1]
Mon, 26 Oct 2026 07:59:20 UTC (3,469 KB)
来源：arXiv cs.CL · arxiv.org