发表机构
University of North Carolina at Chapel Hill; Stevens Institute of Technology; Emory University; Massachusetts Institute of Technology（北卡罗来纳大学教堂山分校; 史蒂文斯理工学院; 埃默里大学; 麻省理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
DrugReason提出多视角推理框架，融合知识图谱与语言模型证据，通过自适应路由和跨专家蒸馏，在多个药物重定位基准上提升预测性能。
AI 中文摘要
药物重定位旨在为现有化合物识别新的治疗用途，与从头药物发现相比，它为临床转化提供了一条更快且更具成本效益的路径。然而，候选药物-疾病对的空间极为庞大，且其潜在关系往往依赖于复杂的多跳生物机制，这使得可靠地预测哪些配对代表真实的治疗关系变得困难。现有方法从两个方向着手解决这一问题：基于知识图谱的方法将整理好的生物医学证据组织成结构化的关系网络，以进行有依据的多跳推理；而基于大语言模型的方法则利用预训练知识生成灵活的机制性解释。然而，单独使用任何一种方法都不够充分——知识图谱局限于已观测到的图结构，而大语言模型缺乏事实依据且存在幻觉风险。为弥补这一差距，我们提出了DrugReason，一个多视角推理框架，它将有依据的知识图谱推理与大语言模型生成的机制性推断相结合，用于药物重定位。DrugReason根据查询上下文自适应地将不同的推理路径路由到专门的专家模块，同时通过跨专家蒸馏目标实现知识共享，而不牺牲专家专业化。在PharmaDB、DDInter和DrugBank上的实验表明，与强单视角推理基线相比，DrugReason提升了平均性能，并且与基于图的替代方法相比取得了具有竞争力或更优的结果，同时提供了可解释的基于路由的预测。
英文摘要
Drug repurposing aims to identify new therapeutic uses for existing compounds and, compared with de novo drug discovery, offers a faster and more cost-effective path to clinical translation. However, the space of candidate drug-disease pairs is enormous and their underlying relationships often depend on complex multi-hop biological mechanisms, making it difficult to reliably predict which pairs represent true therapeutic relationships. Existing approaches tackle this from two directions: knowledge graph-based methods organize curated biomedical evidence into structured relational networks for grounded multi-hop reasoning, while LLM-based methods leverage pretrained knowledge to generate flexible mechanistic rationales. Yet neither is sufficient alone - KGs are confined to observed graph structure while LLMs lack factual grounding and risk hallucination. To address this gap, we propose DrugReason, a multi-view reasoning framework that integrates grounded KG reasoning with LLM-generated mechanistic inference for drug repurposing. DrugReason adaptively routes diverse reasoning paths to specialized experts conditioned on the query context, while a cross-expert distillation objective enables knowledge sharing without sacrificing expert specialization. Experiments on PharmaDB, DDInter, and DrugBank show that DrugReason improves average performance over strong single-view reasoning baselines and achieves competitive or superior results compared with graph-based alternatives, while providing interpretable routing-based predictions.
CommentsEMNLP 2026 Main Conference
