Object Classification on CIFAR-10 (test) (Attack Accuracy, Imp., KNN Dist)
100Attack AccuracyKEDMI + LOM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| KEDMI + LOMDpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 100 | 4.8 | 52.12 | |
| KEDMI + MADpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 100 | 4.8 | 53.17 | |
| KEDMI + LOMMADpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 100 | 4.8 | 63.41 | |
| KEDMIDpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 95.2 | — | 78.24 | |
| GMI + LOMMADpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 95.2 | 52 | 80.3 | |
| GMI + LOMDpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 80.8 | 37.6 | 70.47 | |
| GMI + MADpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 80 | 36.8 | 93.46 | |
| GMIDpub=CIFAR-10, Target Model (Mt)=VGG16, Evaluation Model=ResNet-182023.04 | 43.2 | — | 96.11 |