17 Dec 2024
An algorithmic framework, MedG–KRP, was developed to generate and evaluate medical knowledge graphs from large language models, structured around single concepts. It demonstrates that generalist LLMs show stronger human-rated causal accuracy, while specialized medical LLMs align more closely with structured biomedical ontologies in terms of precision and recall.
There are no more papers matching your filters at the moment.
