Federated Learning Medical Platform MedFL Covers 800 Hospitals Globally: AI Model Training Without Patient Data Leaving Hospitals
The federated learning medical platform MedFL has been deployed across 800 hospitals worldwide, training AI models locally and sharing only encrypted parameter updates.
On August 8, 2028, nonprofit organization HealthAI announced that its federated learning medical platform MedFL has been deployed across 800 hospitals worldwide.
MedFL allows each hospital to train AI diagnostic models locally using its own patient data, then transmit only encrypted model parameter updates (not raw data) to a central server for aggregation. This architecture enables participating hospitals to collectively train a high-quality AI diagnostic system while patient data remains on local servers.
The platform has trained three core models: chest X-ray abnormality detection (97.2% accuracy), skin lesion classification (94.8% accuracy), and pathology slide cancer identification (96.1% accuracy). Each model's training data volume is equivalent to 800 times what a single hospital could accumulate.
Project lead and Stanford professor Li Wei stated that MedFL's success proves the "data stays, models move" approach to medical AI development is viable. Over 200 additional healthcare institutions have applied to join.
Disclaimer
Content is AI-generated. Do not use it as a basis for real decisions. Do not cite it as factual reporting.