Image Classification on CIFAR-10 non-i.i.d. (test) (Accuracy by Rho/Tau Parameters)
88Test Accuracy (rho=0.4, tau=0.3)FedCova
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FedCovanoise_type=asymmetric2026.03 | 88 | 87.84 | 87.77 | 87.83 | 87.31 | 87.29 | |
| FedNoRonoise_type=asymmetric2026.03 | 86.4 | 85.72 | 84.76 | 75.27 | 63.88 | 33.42 | |
| FedNednoise_type=asymmetric2026.03 | 82.3 | 79.73 | 78.54 | 79.77 | 78.03 | 74.67 | |
| RoFLnoise_type=asymmetric2026.03 | 78.87 | 77.85 | 71.56 | 78.45 | 66.1 | 41.38 | |
| FedNed-noise_type=asymmetric2026.03 | 75.69 | 70.38 | 65.51 | 78.63 | 71.22 | 66.44 | |
| FedCorrnoise_type=asymmetric2026.03 | 75.21 | 62.66 | 38.67 | 39.37 | 34 | 30.66 | |
| FedAvgnoise_type=asymmetric2026.03 | 72.06 | 71.23 | 67.4 | 68.91 | 63.53 | 43.4 |