Image Classification on CIFAR-10 (test) (Symmetric/Asymmetric Scores)
95.5Symmetric Acc @ 20%Sel-CL+
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Sel-CL+Backbone=PreAct ResNet-182022.03 | 95.5 | 93.9 | 89.2 | 81.9 | 95.6 | 95.2 | 94.5 | 93.4 | |
| DivideMixBackbone=PreAct ResNet-182022.03 | 95 | 93.7 | 92.4 | 74.2 | 93.8 | 93.2 | 92.5 | 91.4 | |
| ELR (Uns-CL init.)Backbone=PreAct ResNet-18, initialization=Uns-CL2022.03 | 94.4 | 93 | 88.3 | 86.2 | 95 | 94.7 | 94.4 | 93.3 | |
| MOIT+Backbone=PreAct ResNet-182022.03 | 94.1 | 91.8 | 81.1 | 74.7 | 94.2 | 94.3 | 94.3 | 93.3 | |
| M-correctionBackbone=PreAct ResNet-182022.03 | 93.8 | 91.9 | 86.6 | 68.7 | 89.6 | 91.8 | 92.2 | 91.2 | |
| ELRBackbone=PreAct ResNet-182022.03 | 93.8 | 92.6 | 88 | 63.3 | 94.4 | 93.3 | 91.5 | 85.3 | |
| MixupBackbone=PreAct ResNet-182022.03 | 92.3 | 77.6 | 46.7 | 43.9 | 93.3 | 88 | 83.3 | 77.7 | |
| P-correctionBackbone=PreAct ResNet-182022.03 | 92 | 88.7 | 76.5 | 58.2 | 93.1 | 92.9 | 92.6 | 91.6 | |
| GCE (Uns-CL init.)Backbone=PreAct ResNet-18, initialization=Uns-CL2022.03 | 90 | 89.3 | 73.9 | 36.5 | 91.1 | 87.3 | 82.2 | 78.1 | |
| GCEBackbone=PreAct ResNet-182022.03 | 86.6 | 81.9 | 54.6 | 21.2 | 89.5 | 85.6 | 80.6 | 76 | |
| ForwardBackbone=PreAct ResNet-182022.03 | 83.1 | 59.4 | 26.2 | 18.8 | 90.4 | 86.7 | 81.9 | 76.7 | |
| Cross-EntropyBackbone=PreAct ResNet-182022.03 | 82.7 | 57.9 | 26.1 | 16.8 | 88.8 | 86.1 | 81.7 | 76 |