Mitigating Noise Detriment in Differentially Private Federated Learning with Model Pre-training

Open in new window