Abstract: Federated learning (FL) can effectively address the phenomenon of data silos. It allows multiple clients to jointly train machine learning models and protect their local data confidentiality ...