Class-wise Machine Unlearning on CIFAR-20 veh2 (test)
96.08Ar Utility Scorebaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| baselinemodel=ViT2026.02 | 96.08 | 94.35 | 80.96 | |
| MaGAmodel=ViT2026.02 | 95.82 | 0 | 0 | |
| retrainmodel=ViT2026.02 | 95.35 | 0 | 20.36 | |
| AMNCmodel=ViT2026.02 | 94.2 | 0 | 1.26 | |
| UNSIRmodel=ViT2026.02 | 93.83 | 67.29 | 49.96 | |
| SSDmodel=ViT2026.02 | 90.26 | 0 | 9.88 | |
| FTmodel=ViT2026.02 | 85.13 | 0 | 22 | |
| SSDmodel=RN182026.02 | 83.38 | 22.37 | 5.72 | |
| MaGAmodel=RN182026.02 | 82.93 | 0 | 0.08 | |
| baselinemodel=RN182026.02 | 82.84 | 84.99 | 87.72 | |
| retrainmodel=RN182026.02 | 82.41 | 0 | 14.24 | |
| AMNCmodel=RN182026.02 | 82.21 | 0 | 7.84 | |
| UNSIRmodel=RN182026.02 | 81.41 | 58.82 | 45.68 | |
| FTmodel=RN182026.02 | 73.74 | 0 | 40.84 |