Supercharging AI for Biomedical Breakthroughs: Our lab's innovative KG-RAG framework combines knowledge graphs with Large Language Models, boosting AI performance by 71% on complex biomedical questions. Recently accepted in Bioinformatics journal, this groundbreaking approach integrates UCSF's SPOKE biomedical knowledge graph with state-of-the-art AI, democratizing access to vast biomedical knowledge. By providing researchers with more accurate, credible, and transparent answers, KG-RAG paves the way for accelerated discoveries and transformative advancements across the biomedical landscape.
Note: KG-RAG is being integrated into UCSF's Versa generative AI suite, enabling UCSF researchers to leverage this cutting-edge technology in their work.
 

Read the paper!
Check out the code for this paper yourself at our Github repository!