Image Classification on CIFAR-10 (Adversarial Robustness Metrics)
67.6Logit Flipping AccuracyDeepSet-TM
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DeepSet-TMHeterogeneity (α)=0.5, Clients (n)=17, Adversaries (f)=42025.09 | 67.6 | 62.6 | 61 | 59.4 | 51.4 | 48.2 | 48.2 | |
| MeanHeterogeneity (α)=0.5, Clients (n)=17, Adversaries (f)=42025.09 | 65.4 | 59.8 | 58.7 | 55.6 | 42.7 | 24.6 | 24.6 | |
| GMHeterogeneity (α)=0.5, Clients (n)=17, Adversaries (f)=42025.09 | 63.9 | 59.1 | 63.3 | 59.7 | 45.3 | 25.4 | 25.4 | |
| CWTMHeterogeneity (α)=0.5, Clients (n)=17, Adversaries (f)=42025.09 | 63.3 | 59.4 | 62.5 | 59.7 | 44.8 | 27.2 | 27.2 | |
| CWMedHeterogeneity (α)=0.5, Clients (n)=17, Adversaries (f)=42025.09 | 56.7 | 53.3 | 53.8 | 52.3 | 49.3 | 27.8 | 27.8 |