Machine Unlearning on CIFAR-100 (retain)
0.0253Mean KLoMNegGrad (SAM, SAM)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NegGrad (SAM, SAM)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=mid, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0253 | 0.0956 | |
| NegGrad (SAM, SAM)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=high, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0264 | 0.0956 | |
| NegGrad (SAM, SAM)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=low, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0311 | 0.0956 | |
| NegGrad (SGD, SAM)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=mid, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0331 | 0.0956 | |
| NegGrad (SGD, SAM)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=high, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0411 | 0.2097 | |
| NegGrad (SGD, SAM)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SAM, Forget set difficulty (F)=low, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0411 | 0.2097 | |
| NegGrad (SAM, SGD)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=low, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0658 | 0.2097 | |
| NegGrad (SAM, SGD)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=mid, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.066 | 0.2097 | |
| NegGrad (SGD, SGD)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=mid, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0669 | 0.2097 | |
| NegGrad (SAM, SGD)Pre-training Optimizer (A)=SAM, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=high, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0676 | 0.2097 | |
| NegGrad (SGD, SGD)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=high, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0721 | 0.2097 | |
| NegGrad (SGD, SGD)Pre-training Optimizer (A)=SGD, Unlearning Optimizer (U)=SGD, Forget set difficulty (F)=low, Backbone=ResNet50, Unlearning algorithm=NegGrad2025.06 | 0.0747 | 0.2097 |