Class Unlearning on TinyImageNet (TinyIN)
0Average GapGold Std
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gold StdBackbone=ResNet182025.03 | 0 | 0 | 42.15 | |
| LoTUSBackbone=ResNet182025.03 | 0.0925 | 0.0037 | 1.3 | |
| RndLblBackbone=ResNet182025.03 | 0.2925 | 0.0092 | 3.27 | |
| SalUnBackbone=ResNet182025.03 | 0.2925 | 0.0091 | 3.98 | |
| FinetuningBackbone=ResNet182025.03 | 0.2975 | 0.0095 | 3.23 | |
| BadTBackbone=ResNet182025.03 | 0.3125 | 0.0073 | 1.59 | |
| NegGrad+Backbone=ResNet182025.03 | 0.325 | 0.0086 | 3.24 | |
| SSDBackbone=ResNet182025.03 | 0.365 | 0.0035 | 3.19 | |
| SCRUBBackbone=ResNet182025.03 | 0.42 | 0.0073 | 4.05 | |
| UNSIRBackbone=ResNet182025.03 | 0.5075 | 0.0102 | 1.01 |