Abstract: Federated learning (FL) is a decentralized learning paradigm wherein a central server iteratively trains a global model by utilizing clients who possess a certain amount of private datasets.