Federated Learning Robustness on Fashion-MNIST
84.98A_theta Robustness ScoreFedAvg
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| FedAvgSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 84.98 | 12.16 | 55.27 | 7.51 | 7.59 | 74.35 | 6.74 | 73.21 | 73.97 | |
| FoolsGoldSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 84.45 | 3.39 | 4.67 | 3.36 | 2.41 | — | 4.1 | 2.91 | 4.06 | |
| MedianSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 84.39 | 8.83 | 18.29 | 6.07 | 7.82 | 22.43 | 14.87 | 1.53 | 20.7 | |
| M-KrumSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 84.25 | 3.08 | 9.86 | 6.94 | 9.34 | 11.2 | 10.26 | 3.37 | 3.96 | |
| Tr-MeanSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 84.02 | 7.27 | 18.97 | 5.62 | 7.19 | 15.1 | 8.93 | 1.45 | 25.42 | |
| CCSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 83.88 | 2.61 | 19.94 | 6.67 | 9.79 | — | 2.13 | 2.85 | 3.33 | |
| SgnGSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 83.6 | 2.19 | 35.24 | 8.81 | 9.82 | — | 3.75 | 1.83 | 4.15 | |
| FLAMESetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 83.5 | 5.15 | 67.54 | 7.17 | 19.9 | — | 3.94 | 2.56 | 4.03 | |
| FLTrustSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 83.3 | 0.57 | 1.84 | 1.19 | 1.25 | — | 6.27 | 8.32 | 2.67 | |
| FreqFedSetting=Cross-silo, Number of FL clients=40, Global iterations=250, Model Architecture=AlexNet2026.04 | 83.1 | 15.98 | 66.39 | 13.1 | 1.75 | — | 6.11 | 2.43 | 3.24 |