Скачать или смотреть PEPR '25 - Breaking Barriers, Not Privacy: Real-World Split Learning across Healthcare Systems

Breaking Barriers, Not Privacy: Real-World Split Learning across Healthcare Systems

Sravan Kumar Elineni

Regulatory constraints and siloed data often hinder collaborative AI in healthcare. Our project supported by ONC/ASTP implements Split Learning to enable three independent HIEs to jointly train a deep learning model without sharing sensitive patient data. We detail the technical workflow (e.g., partial model hosting, secure exchange of activations) and discuss how we navigated real-world challenges in data integration and quality, network security, and regulatory compliance. Preliminary results show robust model performance and seamless interoperability across participating sites, suggesting a robust blueprint for large-scale privacy-preserving ML in healthcare.

Authors: Sean Muir, Dave Carlson, Himali Saitwal, David E Taylor, Chit win, Jayme Welty, Adam Wong, Jordon Everson, Keith Salzman, Serafina Versaggi, Lindsay Cushing, Savannah Mueller

View the full PEPR 25 program at https://www.usenix.org/conference/pep...