Distributed Backdoor Attack (DBA) Robustness on Federated Learning Dataset
97.58AccuracyMulti-Krum
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Multi-KrumModel=CNN2026.02 | 97.58 | 0.1 | |
| FLTrustModel=CNN2026.02 | 89.31 | 100 | |
| FedAVGModel=CNN2026.02 | 87.95 | 99.9 | |
| FlameModel=CNN2026.02 | 87.89 | 0.4 | |
| Trimmed-MeanModel=VGG192026.02 | 80.44 | 28.92 | |
| MedianModel=VGG192026.02 | 80.13 | 29.77 | |
| FedAVGModel=VGG192026.02 | 78.97 | 25.88 | |
| MedianModel=ResNet182026.02 | 78.87 | 12.16 | |
| Trimmed-MeanModel=ResNet182026.02 | 78.34 | 13.88 | |
| FedAVGModel=ResNet182026.02 | 77.99 | 10.94 | |
| FLTrustModel=ResNet182026.02 | 77.51 | 15.11 | |
| FlameModel=ResNet182026.02 | 75.27 | 3.88 | |
| FLTrustModel=VGG192026.02 | 75.11 | 15.88 | |
| Multi-KrumModel=ResNet182026.02 | 73.02 | 5.61 | |
| Multi-KrumModel=VGG192026.02 | 64.81 | 8.44 | |
| FlameModel=VGG192026.02 | 63.3 | 7.33 |