Large language models encode clinical knowledge
Karan Singhal, Shekoofeh Azizi, Tao Tu et al. · Nature · 2023
Abstract Large language models (LLMs) have demonstrated impressive capabilities, but the bar for clinical applications is high. Attempts to assess the clinical knowledge of models typically rely on automated evaluations based on limited benchmarks. Here, to address these limitations, we present MultiMedQA, a benchmark combining six existing medical question answering datasets spanning professional medicine, research and consumer queries and a new dataset of medical questions searched online, HealthSearchQA. We propose a human evaluation framework for model answers along multiple axes including factuality, comprehension, reasoning, possible harm and bias. In addition, we evaluate Pathways Language Model 1 (PaLM, a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM 2 on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA 3 , MedMCQA 4 , PubMedQA 5 and Measuring Massive Multitask Language Understanding (MMLU) clinical topics 6 ), including 67.6% accuracy on MedQA (US Medical Licensing Exam-style questions), surpassing the prior state of the art by more than 17%. However, human evaluation reveals key gaps. To resolve this, we introduce instruction prompt tuning, a parameter-efficient approach for aligning LLMs to new domains using a few exemplars. The resulting model, Med-PaLM, performs encouragingly, but remains inferior to clinicians. We show that comprehension, knowledge recall and reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Our human evaluations reveal limitations of today’s models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLMs for clinical applications.
Cited by 3,967 works. doi:10.1038/s41586-023-06291-2 Open-access PDF
Authors
- Karan Singhal
- Shekoofeh Azizi
- S. Sara Mahdavi
- Hyung Won Chung
- Ajay Kumar Tanwani
- Heather J. Cole-Lewis
- Perry W. Payne
- Martin G. Seneviratne
- Paul R Gamble
- Christopher J. Kelly
- Abubakr Babiker
- Nathanael Schärli A
- Aakanksha Chowdhery
- Dina Demner‐Fushman
- Blaise Agüera y Arcas
References
- BLEU
- A Learning Algorithm for Boltzmann Machines*
- BioBERT: a pre-trained biomedical language representation model for biomedical text mining
- Discernment and Social Learning as a Companion Training Layer
- BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
- Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
- Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing
- Best Practices for Developing and Validating Scales for Health, Social, and Behavioral Research: A Primer
- Model Cards for Model Reporting
- On the Opportunities and Risks of Foundation Models