Federated Unlearning on CIFAR-100 IID benchmark leaderboard · SOTA2 Research

Federated Unlearning on CIFAR-100 IID

60.67Accuracy (Group [95%, 100%])FedRecovery

Evaluation Results

Method	Links
FedRecovery2026.05		60.67	61.6	57.87	58.4	86.7	83.22	55.01	7.83
FedAU2026.05		59.87	59.73	62.53	60.13	86.6	83.04	53.15	7.26
NoT2026.05		54.8	55.33	53.2	50.93	83.18	79.79	54.72	10.85
FedOSD2026.05		54	47.87	52.53	57.2	85.72	81.21	58.55	7.67
Halimi et al.2026.05		51.47	54.93	52.93	55.2	82.77	78.89	52.56	23.83
Liu et al.2026.05		47.47	43.6	50.8	55.07	87.52	81.99	57.17	7.41
FedMemPrune2026.05		30.8	28.67	39.2	48.53	89.47	80.8	63.33	5.2
Retrained Baseline2026.05		23.2	27.07	36.4	47.6	91.14	81.37	62.29	6.46