Image Classification Efficiency on MNIST (Time to Accuracy Thresholds)
88.6Max AccuracyFAR-SIGN
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FAR-SIGNModel=MLP, Byzantine workers=24%, Clients=512026.05 | 88.6 | 15 | 35 | |
| RFAModel=MLP, Byzantine workers=24%, Clients=512026.05 | 87.5 | 205 | 446 | |
| Multi-KrumModel=MLP, Byzantine workers=24%, Clients=512026.05 | 87.2 | 236 | 599 | |
| MedianModel=MLP, Byzantine workers=24%, Clients=512026.05 | 86.8 | 293 | 542 | |
| KrumModel=MLP, Byzantine workers=24%, Clients=512026.05 | 85.5 | 303 | 867 | |
| BulyanModel=MLP, Byzantine workers=24%, Clients=512026.05 | 85.5 | 321 | 747 | |
| Trimmed MeanModel=MLP, Byzantine workers=24%, Clients=512026.05 | 10.4 | — | — |