Runtime Efficiency (Time to Accuracy) on MNIST (Image Classification)
11Time to 80% Acc (s)FAR-SIGN
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FAR-SIGNModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 11 | 24 | 89.6 | |
| RFAModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 223 | 417 | 87.5 | |
| MedianModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 247 | 467 | 87.6 | |
| Trimmed MeanModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 260 | 511 | 87.5 | |
| Multi-KrumModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 288 | 594 | 86.9 | |
| KrumModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 314 | 803 | 85.8 | |
| BulyanModel=MLP, Byzantine workers=24%, Clients=51, Seeds=52026.05 | 374 | 772 | 85.8 |