Fed-BioMed
InriaCollaborative learning framework for biomedical research from Inria and Université Côte d'Azur. Hospitals run Fed-BioMed nodes that hold and manage their datasets through a node GUI, and a researcher component sends PyTorch or scikit-learn training plans to them for federated learning or federated analytics, with Joye-Libert or low-overhead masking secure aggregation.
Openness
5 high confidence- license
- Apache-2.0(OSI
- source
- public(node, researcher and GUI components)
- core features withheld
- no — an Inria and UCA academic project with no paid edition
Fed-BioMed is Apache-2.0, even though GitHub labels it "other": the license file is a short notice that grants Apache-2.0 by SPDX identifier. As an academic project from Inria and Université Côte d'Azur, it has no paid edition to hold anything back for.
- https://raw.githubusercontent.com/fedbiomed/fedbiomed/HEAD/LICENSE.md recorded 2026-09-27
LICENSE.md: "originally developed by Inria ... and of Université Côte d'Azur (UCA)", "Licensed under the Apache License, Version 2.0", "SPDX-License-Identifier: Apache-2.0".
- https://raw.githubusercontent.com/fedbiomed/fedbiomed/HEAD/README.md recorded 2026-09-27
README: "Fed-BioMed is an open source project focused on empowering healthcare using non-centralized approaches"; installs as fedbiomed[node, gui, researcher]. No paid tier is named.
- https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/fedbiomed%2Ffedbiomed recorded 2026-09-27
Repository record for fedbiomed/fedbiomed: license "other" (the classifier's reading of the short notice), not archived, not a fork, pushed 2026-09-22, 93 stars.
- https://ungh.cc/repos/fedbiomed/fedbiomed/files/develop recorded 2026-09-27
The 912-file tree of fedbiomed/fedbiomed at `develop`: one root LICENSE, and no ee/, enterprise/, commercial/, proprietary/ or premium/ path anywhere in it. The license file is LICENSE.md. Path absence only, paired with the statement cited beside it.
Adoption
1 medium confidencePyPI downloads of `fedbiomed`, the package that installs the node, researcher and GUI components. Hospital deployments also run from the project's Docker images, which this does not count.
- https://pypistats.org/api/packages/fedbiomed/recent recorded 2026-09-27
last_month = 83 downloads for the package `fedbiomed`.
Capability
4 medium confidenceEach hospital runs its own node and manages its datasets through a GUI, and model parameters are encrypted before aggregation. That matches Flower's production reach.
- https://docs.fedbiomed.org/latest/user-guide/secagg/introduction/ recorded 2026-09-27
Secure aggregation: local parameters "are encrypted before sending them to the researcher"; two schemes, Joye-Libert ("an additive homomorphic encryption scheme") and Low-Overhead Masking; in Joye-Libert "each party manually registers the certificates of other parties".
- https://raw.githubusercontent.com/fedbiomed/fedbiomed/HEAD/README.md recorded 2026-09-27
README: `fedbiomed node start`, `fedbiomed node dataset add`, and a Node GUI "to manage datasets and deploy new ones".
- https://ungh.cc/repos/fedbiomed/fedbiomed/files/develop recorded 2026-09-27
The tree ships docker/node/Dockerfile and docker/researcher/Dockerfile, and a certificate_manager module.
Verified 2026-09-27