Fed-BioMed
A self-hosted federated learning platform for healthcare research that keeps patient data at each institution and works with PyTorch and Scikit-learn.
Fed-BioMed lets hospitals and researchers train AI models across institutions while patient data stays on each institution's own infrastructure. It's a self-hosted platform for biomedical research that needs access to data across multiple sites but faces restrictions on sharing clinical records.
Participating institutions control access to their resources and approve every training activity. They can also monitor research activity, so collaboration doesn't require handing all authority to a central research team. A federated network connects the sites while each provider retains control of its patient data.
The platform covers work before model training, too. Shared metadata helps researchers discover which data is available across institutions, and federated analytics lets them explore distributed datasets without collecting those datasets in one place. Researchers can use these capabilities for collaborative experiments in real clinical environments.
Fed-BioMed uses Python, PyTorch and Scikit-learn, making it relevant to teams already working with those machine learning libraries. It runs on Linux and macOS, with Windows support through WSL2. Separate node and researcher components serve the institutions hosting data and the teams conducting experiments. The platform is used in multi-center hospital collaborations and European research projects, including oncology work with Unicancer hospital partners.
