Robust Image Classification on SVHN (test) benchmark leaderboard · SOTA2 Research

Robust Image Classification on SVHN (test)

95.84Clean AccuracyARNAS

Evaluation Results

Method
ARNAS2024.05	95.84	94.43	—	—	—	—	—	92.02
PDARTS2024.05	95.1	93.01	—	—	—	—	—	89.58
Prototype Conformity Loss (PCL)2019.04	94.4	87.1	82.2	80.7	—	—	—	—
ResNet-182024.05	92.06	88.73	—	—	—	—	—	69.51
Oracle2022.11	91.95	—	—	53.89	50.59	47.05	—	—
Two-stage2022.11	90.62	—	—	53.92	50.86	47.31	—	—
Warm-up+PLA2022.11	90.5	—	—	54.58	51.04	47.47	—	—
Song et al.2019.04	82.9	57.2	53.9	53.2	—	—	—	—
CalFAT2022.05	44.57	17.63	—	—	15.37	11.49	15.21	15.6
FedGAIRAT2022.05	34.92	16.18	—	—	11.34	10.87	14.68	14.11
MixFAT2022.05	34.43	15.69	—	—	11.31	9.06	14.6	14.36
FedPGD2022.05	33.96	16.07	—	—	10.87	9.53	14.67	14.3
FedTRADES2022.05	29.55	15.01	—	—	10.58	9.71	12.83	12.01
FedRBN2022.05	28.55	14.69	—	—	11.42	9.41	12.91	12.3
FedMART2022.05	19.96	13	—	—	9.92	8.83	11.8	11.49
EXP2022.11	19.6	—	—	19.6	19.56	19.65	—	—
FORWARD2022.11	19.59	—	—	19.59	19.59	19.68	—	—
FREE2022.11	19.59	—	—	19.59	19.59	19.68	—	—
NN2022.11	19.59	—	—	19.59	19.59	19.68	—	—
SCL_NL2022.11	19.59	—	—	19.59	19.58	19.67	—	—
SCL_EXP2022.11	19.59	—	—	19.59	19.58	19.68	—	—
LOG2022.11	19.59	—	—	19.6	19.59	19.68	—	—