Image Classification on CIFAR10 (Test Accuracy and Communication Efficiency)
83.2AccuracyQSGD[2]+FedBCGD+
Evaluation results
| Method | Links | ||
|---|---|---|---|
| QSGD[2]+FedBCGD+Backbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 83.2 | 29 | |
| QSGD[2]+FedBCGDBackbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 82.6 | 32 | |
| FedBCGD+Backbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 80.6 | 57 | |
| FedBCGDBackbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 77.2 | 65 | |
| FedPAQBackbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 75.2 | 121 | |
| TOP-kBackbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 74.5 | 92 | |
| FedAvgBackbone=LeNet-5, Heterogeneity (ρ)=0.6, Number of clients (N)=5, Total communication floats budget=200d2026.03 | 73.2 | 133 |