On Joint Noise Scaling in Differentially Private Federated Learning with Multiple Local Steps

Open in new window