Think and Retrieval: A Hypothesis Knowledge Graph Enhanced Medical Large Language Models

Xinke Jiang,Ruizhe Zhang,Yongxin Xu,Ruolun Qiu,Yuehui Fang,Zhiyuan Wang,Jialei Tang,Hongxin Ding,Chunyan Xu,Jun Zhao,Yasha Wang
DOI: https://doi.org/10.48550/arxiv.2312.15883
2023-01-01
Abstract:We explore how the rise of Large Language Models (LLMs) significantly impacts task performance in the field of Natural Language Processing. We focus on two strategies, Retrieval-Augmented Generation (RAG) and Fine-Tuning (FT), and propose the Hypothesis Knowledge Graph Enhanced (HyKGE) framework, leveraging a knowledge graph to enhance medical LLMs. By integrating LLMs and knowledge graphs, HyKGE demonstrates superior performance in addressing accuracy and interpretability challenges, presenting potential applications in the medical domain. Our evaluations using real-world datasets highlight HyKGE's superiority in providing accurate knowledge with precise confidence, particularly in complex and difficult scenarios. The code will be available until published.
What problem does this paper attempt to address?