Sub-class Machine Unlearning on CIFAR-20 Rocket sub-class
85.26RAPretrain
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PretrainBackbone=ResNet-182024.02 | 85.26 | 80.73 | 85.21 | 26.53 | 92.89 | 6,910 | |
| BTBackbone=ResNet-182024.02 | 85.24 | 2.8 | 84.36 | 0.26 | 0 | 27 | |
| RetrainBackbone=ResNet-182024.02 | 84.85 | 2.69 | 84.07 | 0 | 12.06 | 4,298 | |
| SalUnBackbone=ResNet-182024.02 | 84.82 | 2.99 | 84 | 0.13 | 0 | 1,042 | |
| SCRUBBackbone=ResNet-182024.02 | 84.73 | 5.8 | 83.84 | 1.15 | 13.28 | 113 | |
| MU-MisBackbone=ResNet-182024.02 | 84.28 | 2.91 | 83.5 | 0.49 | 0.07 | 21 | |
| SSDBackbone=ResNet-182024.02 | 84.23 | 2.6 | 83.35 | 0.48 | 3.76 | 18 | |
| FTBackbone=ResNet-182024.02 | 82.7 | 4.2 | 81.97 | 1.92 | 5.4 | 138 | |
| DUCKBackbone=ResNet-182024.02 | 82.09 | 19.4 | 81.43 | 7.37 | 32.84 | 58 | |
| MUNBaBackbone=ResNet-182024.02 | 81.43 | 7 | 80.8 | 3.67 | 7.2 | 362 | |
| SCARBackbone=ResNet-182024.02 | 76.49 | 43.81 | 76.26 | 19.09 | 28.04 | 442 | |
| NGBackbone=ResNet-182024.02 | 62.84 | 5.67 | 62.48 | 15.52 | 72.7 | 4 | |
| RLBackbone=ResNet-182024.02 | 60.89 | 6.52 | 60.5 | 17.11 | 3.7 | 5 | |
| JiTBackbone=ResNet-182024.02 | 59.15 | 4 | 58.6 | 17.49 | 29.03 | 4 | |
| LoTusBackbone=ResNet-182024.02 | 35.93 | 39 | 36.04 | 44.42 | 18.6 | 105 |