The Fed-BioMed Team is a multidisciplinary group of researchers and engineers working on privacy-preserving artificial intelligence for healthcare. The project is developed in collaboration with academic institutions, hospitals, and industry partners to enable secure and trustworthy collaborative learning.
Fed-BioMed provides an open-source framework that allows multiple healthcare institutions to train machine learning models together without sharing sensitive patient data. Instead of moving data, models are trained locally at each institution and only aggregated parameters are shared, preserving privacy and regulatory compliance.
The initiative focuses on building practical federated learning tools for real-world medical environments, including hospitals and research networks. The platform supports several machine learning frameworks and integrates security mechanisms such as secure aggregation and privacy-preserving protocols.
Fed-BioMed is actively used in collaborative research projects across Europe, particularly in oncology and medical imaging, and continues to evolve through contributions from an international community of researchers, clinicians, and developers.
