Large Language Models ↔ Medical Knowledge Graph

Relations (1)

related 0.80 — strongly supporting 8 facts

Large Language Models are utilized to evaluate the accuracy and usability of the Medical Knowledge Graph [1], [2], [3], while the graph itself serves as a structured knowledge source to improve the correctness and reasoning capabilities of these models [4], [5], [6].

Facts (8)

Sources

Bridging the Gap Between LLMs and Evolving Medical Knowledge arxiv.org arXiv 6 facts

claimThe Medical Knowledge Graph (MKG) is designed to be both human-readable and usable by advanced LLMs, serving as a tool for medical QA and decision-making.

procedureThe evaluation of the Medical Knowledge Graph (MKG) involved a two-phase process using expert LLMs specialized in medical domains to assess accuracy, robustness, and usability.

measurementIn the second phase of MKG evaluation, expert LLMs achieved an 89% accuracy rate when answering complex medical queries requiring multi-hop reasoning, such as managing comorbidities or determining multi-drug treatment protocols.

measurementIn the first phase of MKG evaluation, expert LLMs independently rated graph components on a scale of 1 to 10, resulting in an average accuracy score of 8.9/10 for node identification, 8.8/10 for relationship relevance, and 8.5/10 for the clarity and precision of node summaries.

measurementThe integration of LLMs with medical knowledge graphs for HIV/AIDS queries achieved a 9.4/10 rating for interpretability, provided contextually accurate responses regarding drug interactions and side effects, and received a 10/10 rating for relevance and accuracy.

measurementThe automatically constructed medical knowledge graphs described in 'Bridging the Gap Between LLMs and Evolving Medical Knowledge' contain approximately 76,681 nodes and 354,299 edges.

A survey on augmenting knowledge graphs (KGs) with large ... link.springer.com Springer 1 fact

claimIntegrating a medical knowledge graph is a method to ensure correct diagnoses and treatment options generated by Large Language Models.

Large Language Models Meet Knowledge Graphs for Question ... arxiv.org arXiv 1 fact

referenceFact Finder, developed by Fraunhofer IAIS and Bayer, augments Large Language Models with query-based retrieval from medical knowledge graphs to improve the completeness and correctness of generated answers.