ALI-DPFL: Differentially Private Federated Learning with Adaptive Local Iterations

arXiv:2308.10457v4 Announce Type: replace-cross
Abstract: Federated Learning (FL) is a distributed machine learning technique that allows model training among multiple devices or organizations by sharing training parameters instead of raw data. However, adversaries can still infer individual information through inference attacks (e.g. differential attacks) on these training parameters. As a result, Differential Privacy (DP) has been widely used in FL to prevent such attacks.
We consider differentially private federated learning in a resource-constrained scenario, where both privacy budget and communication …

adversaries arxiv attacks can cs.cr cs.lg data devices distributed federated federated learning information local machine machine learning model training organizations private result sharing training