Image Classification on CIFAR-10 (Target Accuracy Rounds)
86.52Accuracy (500 Rounds)FedDC‡
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FedDC‡Federated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 86.52 | 87.47 | 323 | 519 | |
| FedACGFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 85.13 | 89.1 | 319 | 450 | |
| FedDynFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 84.82 | 88.1 | 392 | 646 | |
| MOONFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 83.32 | 86.3 | 371 | 686 | |
| FedAvgMFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 80.56 | 85.48 | 519 | 828 | |
| FedCM†Federated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 78.92 | 83.71 | 624 | 1,000 | |
| FedLCFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 78.37 | 84.79 | 680 | 1,000 | |
| FedDecorrFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 76.62 | 83.4 | 728 | 1,000 | |
| FedNTDFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 76.05 | 83.78 | 685 | 1,000 | |
| FedMLBFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 74.98 | 84.04 | 714 | 1,000 | |
| FedAvgFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 74.36 | 82.53 | 840 | 1,000 | |
| FedProxFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 73.7 | 82.68 | 826 | 1,000 | |
| FedADAMFederated Learning Setting=Moderate-scale, Number of clients=100, Participation rate=5%, Dirichlet parameter=0.32022.01 | 72.33 | 81.73 | 908 | 1,000 |