List your company
ResearchFormerly WizWand

Image Classification on Fashion-MNIST Sample Unlearning (Df)

100AccuracyOriginal
-4235077Nov 13, 2024
Updated 5mo ago

Evaluation results

MethodLinks
OriginalBackbone=ResNet182024.11100
FinetuneBackbone=ResNet182024.11100
SCRUBBackbone=ResNet182024.11100
FinetuneBackbone=ViT2024.11100
OriginalBackbone=ResNet182024.11100
OriginalBackbone=ViT2024.1199.22
BadTBackbone=ViT2024.1199.22
SCRUBBackbone=ViT2024.1199.22
Residual Feature Alignment UnlearningBackbone=ViT2024.1199.22
OriginalBackbone=ViT2024.1198.9
RetrainBackbone=ResNet182024.1197.66
Residual Feature Alignment UnlearningBackbone=ResNet182024.1197.66
NegGradBackbone=ResNet182024.1196.88
NegGradBackbone=ViT2024.1196.88
RetrainBackbone=ViT2024.1196.09
FinetuneBackbone=ResNet182024.1194.53
BadTBackbone=ResNet182024.1189.84
FinetuneBackbone=ViT2024.1167.02
NegGradBackbone=ResNet182024.119.87
SCRUBBackbone=ResNet182024.116.75
Residual Feature Alignment UnlearningBackbone=ResNet182024.110.63
BadTBackbone=ViT2024.110.58
SCRUBBackbone=ViT2024.110.58
Residual Feature Alignment UnlearningBackbone=ViT2024.110.07
BadTBackbone=ResNet182024.110.02
RetrainBackbone=ResNet182024.110
RetrainBackbone=ViT2024.110
NegGradBackbone=ViT2024.110