Federated learning is regarded as a viable method for implementing innovative healthcare applications without exchanging raw data.
CERN has successfully designed, developed, and implemented a Federated Learning (FL) Platform through the CAFEIN Project, financed by the MA budget. 
A final step towards fully deploying a patient-centred solution is the development of global AI algorithms trained on IoT edge devices via a federated training platform.
The project aims to design, develop, and deploy AI algorithms that can run directly on IoT devices (to avoid the need of any intermediate data storage, thus further preserving patients’ privacy and data control) with the use case of the TRUSTroke project, to support the monitoring, diagnosis and therapy evaluation of stroke patients based on the achieved developments and synergies with the ongoing project, collaborations, datasets and clinician support.