List your company
ResearchFormerly WizWand

Image Classification on MNIST (test) (Unlearning Metrics)

77.48Pretrain Accuracyf-FUM (KL-KL)
18.137633.543848.9564.3562Feb 5, 2026
Updated 8mo ago

Evaluation results

MethodLinks
f-FUM (KL-KL)Confuse=10, Removed (%)=62026.0277.4894.87
f-FUM (KL-JS)Confuse=10, Removed (%)=62026.0277.4898.09
f-FUM (KL-χ²)Confuse=10, Removed (%)=62026.0277.4893.21
NoTConfuse=10, Removed (%)=62026.0277.4873.52
HalimiConfuse=10, Removed (%)=62026.0277.4895.27
f-FUM (KL-KL)Confuse=5, Removed (%)=102026.0224.9991.39
f-FUM (KL-JS)Confuse=5, Removed (%)=102026.0224.9997.49
f-FUM (KL-χ²)Confuse=5, Removed (%)=102026.0224.9985.9
NoTConfuse=5, Removed (%)=102026.0224.999.87
HalimiConfuse=5, Removed (%)=102026.0224.9992.26
f-FUM (KL-KL)Confuse=5, Removed (%)=62026.0220.4298.8
f-FUM (KL-JS)Confuse=5, Removed (%)=62026.0220.4297.78
f-FUM (KL-χ²)Confuse=5, Removed (%)=62026.0220.4282.82
NoTConfuse=5, Removed (%)=62026.0220.429.78
HalimiConfuse=5, Removed (%)=62026.0220.4298.61