Peer-Reviewed Publication
NPJ Digit Med2025;8(1):429.July 11, 2025Journal Article
A perspective for adapting generalist AI to specialized medical AI applications and their challenges.
Zifeng Wang1, Hanyin Wang2,3, Benjamin Danek1, Ying Li4, Christina Mack5, Luk Arbuckle5, Devyani Biswal5, Hoifung Poon6, Yajuan Wang7, Pranav Rajpurkar8, Cao Xiao9, Jimeng Sun10,11,12
1Keiji AI, Seattle, USA.
2Department of Computer Science, University of Illinois Urbana-Champaign, Urbana, IL, USA.
3Division of Hospital Internal Medicine, Mayo Clinic Health System, Mankato, MN, USA.
4Regeneron, Tarrytown, NY, USA.
5IQVIA, Durham, NC, USA.
6Microsoft Research, Redmond, WA, USA.
7Teladoc Health, Harrison, NY, USA.
8Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA.
9GE Healthcare, Bellevue, WA, USA.
10Keiji AI, Seattle, USA. jimeng@illinois.edu.
11Department of Computer Science, University of Illinois Urbana-Champaign, Urbana, IL, USA. jimeng@illinois.edu.
12Carle Illinois College of Medicine, University of Illinois Urbana-Champaign, Urbana, IL, USA. jimeng@illinois.edu.
Abstract
We introduce a framework to adapt large language models for medicine: (1) Modeling: breaking down medical workflows into manageable steps; (2) Optimization: optimizing model performance via advanced adaptations; and (3) System engineering: developing agent or chain systems. Furthermore, we describe varied use cases, such as clinical trial design, clinical decision support, and medical imaging anal…
Create a free account to keep reading
Free members get 10 full research views every month across publications, clinical trials, FDA clearances, adverse events, and NIH grants. No credit card required.
Want unlimited research access? See Pro plans