Selective Unlearning on CIFAR-100 Selective class-0 examples benchmark leaderboard · SOTA2 Research

Selective Unlearning on CIFAR-100 Selective class-0 examples

37.46Test Error (Mean)Original

Evaluation Results

Method	Links
Original2023.02		37.46	4.99	54	174	0	0	58	9.1
Original2023.02		37.46	4.99	54	174	0	0	63	10.96
Retrain2023.02		37.54	5.24	57	184	19	15.81	45	19.04
Retrain2023.02		37.74	5.31	56	177	19.5	12	49	11.47
SCRUB2023.02		38.18	4.78	48	144	45	13.6	56	19.76
SCRUB2023.02		38.29	4.61	49	142	29	16.28	54.5	18.54
SCRUB+R2023.02		38.39	4.63	68	200	23	8.7	49.5	11.68
SCRUB+R2023.02		38.4	4.7	69	207	40	13.33	61	17.59
NegGrad+2023.02		45.56	7	1,114	1,257	47	9.89	67.5	13.94
Bad-T@epoch102023.02		45.88	5.7	1,203	944	92	10.25	70	10.87
NegGrad+2023.02		45.98	6.79	1,164	1,202	44	11.15	55	23.88
Bad-T@epoch12023.02		46.46	6.28	1,295	1,171	55.5	16.2	55	21.59
Bad-T@epoch102023.02		47.09	5.65	1,380	979	96	8.07	77	17.25
Bad-T@epoch12023.02		47.23	5.76	1,379	1,028	66	18.22	62	30.2