Image Classification on CIFAR10 Non-IID (Dirichlet alpha = 0.8)
0.6881AccuracyFLEX-MoE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FLEX-MoEIndicator=Loss-driven, Client count C=20, Expert count E=8, Expert selection range kc=[2, 6]2025.12 | 0.6881 | 0.003 | 14,085 | |
| FLEX-MoEIndicator=Acc-driven, Client count C=20, Expert count E=8, Expert selection range kc=[2, 6]2025.12 | 0.6452 | 0.003 | 14,564 | |
| GreedyIndicator=Acc-driven, Client count C=20, Expert count E=8, Expert selection range kc=[2, 6]2025.12 | 0.6226 | 0.2092 | 1,178,539 | |
| GreedyIndicator=Loss-driven, Client count C=20, Expert count E=8, Expert selection range kc=[2, 6]2025.12 | 0.6114 | 0.2951 | 1,794,908 | |
| RandomIndicator=None, Client count C=20, Expert count E=8, Expert selection range kc=[2, 6]2025.12 | 0.567 | 0.0283 | 202,365 |