Peer-Reviewed Publication
J Biomed Inform2026;105101.September 17, 2026Journal Article
MEDAL: Sequential adapter learning for privacy-preserving multicenter clinical language models.
Ahmed Bakr1, Augusto Garcia-Agundez2, Travis Atkison3, Vivek A Rudrapatna4, Trinabh Gupta5
1Department of Computer Science, The University of Alabama, Tuscaloosa, AL, 35401, USA; DataUnite, Inc., Cupertino, CA, 95014, USA.
2Division of Rheumatology, University of California San Francisco, San Francisco, CA, 94143, USA.
3Department of Computer Science, The University of Alabama, Tuscaloosa, AL, 35401, USA.
4Bakar Computational Health Sciences Institute, University of California, San Francisco, CA, 94143, USA; Division of Gastroenterology, School of Medicine, University of California, San Francisco, CA, 94143, USA. Electronic address: trinabh@ucsb.edu.
5DataUnite, Inc., Cupertino, CA, 95014, USA; Department of Computer Science, University of California, Santa Barbara, CA, 93106, USA. Electronic address: vivek.rudrapatna@ucsf.edu.
Abstract
BACKGROUND: Scalable reasoning over clinical notes is a major goal in medical artificial intelligence. Large Language Models (LLMs) are promising for this task, particularly when fine-tuned on multicenter note corpora, but progress has been limited by barriers to data sharing and technical gaps in Federated Learning (FL).
METHODS: We developed Multi-institutional Efficient/Distributed Adapter Lea…
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