Random Unlearning on CIFAR-10 (original)
98.81ArMaGA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MaGAbackbone=ViT2026.02 | 98.81 | 96.27 | 29.18 | |
| retrainbackbone=ViT2026.02 | 98.75 | 99.41 | 91.67 | |
| baselinebackbone=ViT2026.02 | 98.66 | 99.71 | 88.66 | |
| SSDbackbone=ViT2026.02 | 98.63 | 99.39 | 88.22 | |
| Bad Teacherbackbone=ViT2026.02 | 98.17 | 93.52 | 26.8 | |
| Amnesiacbackbone=ViT2026.02 | 98.15 | 74.88 | 6.4 | |
| FTbackbone=ViT2026.02 | 97.64 | 98.24 | 86.83 | |
| baselinebackbone=ResNet-182026.02 | 95.31 | 94.03 | 76.53 | |
| MaGAbackbone=ResNet-182026.02 | 93.04 | 85.59 | 34.04 | |
| retrainbackbone=ResNet-182026.02 | 92.42 | 94.4 | 74.42 | |
| SSDbackbone=ResNet-182026.02 | 91.38 | 95.88 | 76.5 | |
| Bad Teacherbackbone=ResNet-182026.02 | 90.68 | 93.13 | 47.08 | |
| Amnesiacbackbone=ResNet-182026.02 | 90.63 | 55.35 | 17.89 | |
| FTbackbone=ResNet-182026.02 | 87.86 | 88.77 | 71.7 |