KGA
Knowledge Graph Agent—the named methodology developed by Harvard researchers that integrates large language models with knowledge graphs through a four-phase process (generate, review, revise, answer) to improve medical question-answering
As defined in Harvard Presents NEW Knowledge-Graph AGENT (MedAI)
Where it comes up · 1
Most important claims · 3 of 3Search all claims for “KGA” →
- On multiple-choice medical questions with Llama 3.8B, applying the KGA methodology improves accuracy from approximately 50% to 70% or higher, with larger improvements on questions with more medical concepts.Harvard Presents NEW Knowledge-Graph AGENT (MedAI)
- The methodology is called 'Knowledge Graph Agent' or 'KGA' (per the Harvard publication), though the speaker notes uncertainty about what the acronym fully stands for.Harvard Presents NEW Knowledge-Graph AGENT (MedAI)
- The research was conducted using a cluster of 4 Nvidia H100 GPUs, demonstrating that the computational requirements for implementing the KGA methodology are not prohibitively high.Harvard Presents NEW Knowledge-Graph AGENT (MedAI)