Image Classification on CIFAR-10 Distribution II: Quantity Skew (test)
45.16AccuracyNo Protection
Evaluation results
| Method | Links | ||
|---|---|---|---|
| No ProtectionBackbone=ResNet-182026.01 | 45.16 | — | |
| LaDP-FLepsilon=0.5, Backbone=ResNet-182026.01 | 31.79 | 108,364 | |
| AdapLDPepsilon=0.5, Backbone=ResNet-182026.01 | 29.52 | 229,980 | |
| DPA LDPepsilon=0.5, Backbone=ResNet-182026.01 | 27.91 | 140,293 | |
| LaDP-FLepsilon=0.2, Backbone=ResNet-182026.01 | 26.86 | 274,003 | |
| Sensitive DPepsilon=0.5, Backbone=ResNet-182026.01 | 26.03 | 112,989 | |
| DPA LDPepsilon=0.2, Backbone=ResNet-182026.01 | 24.95 | 398,153 | |
| Sensitive DPepsilon=0.2, Backbone=ResNet-182026.01 | 21.39 | 283,316 | |
| Time-Varying DPepsilon=0.5, Backbone=ResNet-182026.01 | 20.94 | 162,628 | |
| AdapLDPepsilon=0.2, Backbone=ResNet-182026.01 | 19.1 | 700,645 | |
| Time-Varying DPepsilon=0.2, Backbone=ResNet-182026.01 | 17.77 | 356,558 | |
| Full DPepsilon=0.5, Backbone=ResNet-182026.01 | 17.33 | 211,161 | |
| Full DPepsilon=0.2, Backbone=ResNet-182026.01 | 15.13 | 727,895 |