git clone https://github.com/pliang279/LG-FedAvg.git Data We run FedAvg and LG-FedAvg experiments on MNIST (link) and CIFAR10 (link). See our paper for a description how we process and partition the data for federated learning experiments. ...
We propose CLDP-pFedAvg, a novel client-level differentially private federated meta-learning scheme, which guarantees differential privacy on a client-by-client basis for federated learning with heterogeneous clients. CLDP-pFedAvg is based on FedAvg, the most popular learning algorithm for federated le...
Additionally, we perform detailed analyses (e.g., convergence analysis) to demonstrate CLDP-pFedAvg’s convergence rate. To the best of our knowledge, our approach is the first to combine a meta-based FedAvg algorithm with client-level differential privacy in federated learning. In summary, the...