Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Single-class unlearning on STL-10
Loading...
89.41
FA (Forgetting Accuracy)
Original
-3.3476
20.7337
44.815
68.8963
Apr 13, 2026
FA (Forgetting Accuracy)
RA (Retention Accuracy)
ERB (Error Bound/Score)
IL (Influence Loss)
MIA (Membership Inference Attack Success Rate)
Updated 3d ago
Evaluation Results
Method
Method
Links
FA (Forgetting Accuracy)
RA (Retention Accuracy)
ERB (Error Bound/Score)
IL (Influence Loss)
MIA (Membership Inference Attack Success Rate)
Original
Backbone=ViT-L/16
2026.04
89.41
89.12
18.93
-
-
GPPU
Backbone=ResNet-50
2026.04
0.29
90.34
94.68
80
57.9
GPPU
Backbone=ViT-L/16
2026.04
0.22
91.14
95.84
60
57.1
Feedback
Search any
task
Search any
task
