检索说明

  • 查询:GraphRAG evidence retrieval
  • 年份过滤:2025-
  • 单查询上限:20
  • 发现日期:2026-11-12

Semantic Scholar 候选元数据

  • 作者:Xuan-Wen Liu
  • 年份:2026
  • Venue:Proceedings of the 2nd International Conference on Software Engineering and Computer Applications
  • DOI:10.1145/3829641.3829712
  • arXiv:未返回
  • S2 citationCount(当前快照):0
  • 来源线索:https://doi.org/10.1145/3829641.3829712

摘要(未核验)

Clinical deployment of Large Language Models (LLMs) is severely bottle-necked by factual hallucinations, while traditional Knowledge Graph Question Answering (KGQA) architectures are constrained by rigid templates and limited multi-hop reasoning. To resolve this, we formulate a novel Graph-enhanced Retrieval-Augmented Generation (GraphRAG) framework for reliable, interpretable medical QA. Methodologically, we construct a massive medical knowledge graph via Open Information Extraction (OpenIE) and integrate Graph Neural Networks (GNNs) to encode topological structures, enabling complex latent reasoning. The retrieved subgraphs subsequently act as factual anchors to augment the LLM, systematically mitigating hallucination risks. Coupled with an evidence provenance mechanism for verifiability, our framework empirically outperforms baseline LLMs and KGQA systems across accuracy, reasoning depth, and factual fidelity.

转为正式文献前的核验清单

  • [ ] 从论文、出版社、会议、ACL Anthology、OpenReview 或 arXiv 页面核对标题、作者、年份和版本。
  • [ ] 核对 DOI、venue、正式链接和是否存在撤稿/版本变化。
  • [ ] 阅读正文后补充方法、证据边界、结果和与现有研究的关系。

Semantic Scholar 记录用于发现候选,不是正式一手引文,也不支持对结果覆盖范围或研究质量的穷尽性判断。