FEderated Learning with moDel ownerShip protection and Privacy ARmoring (FELDSPAR)

Advances in AI are driving digital transformation but rely on access to vast, often privacy-sensitive datasets distributed across institutions constrained by regulations like the GDPR. Federated Learning (FL) offers a privacy-preserving approach by enabling collaborative model training without sharing raw data, though it still faces risks such as data reconstruction, property or membership inference attacks, and copyright or ownership issues due to the distributed nature of model sharing.