Misclassification Detection on CIFAR-10
99.8AUROCLSH-Shapley
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LSH-ShapleyN=50,000, d=7682026.05 | 99.8 | — | — | — | — | — | |
| KNNN=50,000, d=7682026.05 | 99.5 | — | — | — | — | — | |
| OpenMixNetwork=WRNet2023.03 | 94.81 | 2.32 | 22.08 | 97.16 | — | — | |
| OpenMixNetwork=DenseNet2023.03 | 93.57 | 4.68 | 33.57 | 95.51 | — | — | |
| DoctorNetwork=DenseNet2023.03 | 93.19 | 5.64 | 38.29 | 94.78 | — | — | |
| MSPNetwork=WRNet2023.03 | 93.14 | 4.76 | 30.15 | 95.91 | — | — | |
| MSPNetwork=DenseNet2023.03 | 93.14 | 5.66 | 38.64 | 94.78 | — | — | |
| DoctorNetwork=WRNet2023.03 | 93.13 | 4.75 | 30.46 | 95.91 | — | — | |
| RegMixupNetwork=WRNet2023.03 | 92.31 | 3.36 | 37.48 | 97.1 | — | — | |
| SSNetwork=VGG-162023.03 | 92.22 | — | 44.69 | — | — | — | |
| OpenMixNetwork=ResNet1102023.03 | 92.09 | 6.31 | 39.63 | 94.98 | — | — | |
| RegMixupNetwork=DenseNet2023.03 | 92.02 | 5.2 | 41.5 | 95.5 | — | — | |
| LogitNormNetwork=WRNet2023.03 | 91.06 | 5.81 | 46.06 | 95.5 | — | — | |
| MixupNetwork=WRNet2023.03 | 90.79 | 5.3 | 29.68 | 96.71 | — | — | |
| MaxLogitNetwork=WRNet2023.03 | 90.6 | 6.85 | 37.01 | 95.91 | — | — | |
| EnergyNetwork=WRNet2023.03 | 90.47 | 6.91 | 39.13 | 95.91 | — | — | |
| DoctorNetwork=ResNet1102023.03 | 90.15 | 9.51 | 42.95 | 94.3 | — | — | |
| MSPNetwork=ResNet1102023.03 | 90.13 | 9.52 | 43.33 | 94.3 | — | — | |
| MixupNetwork=DenseNet2023.03 | 89.87 | 9.55 | 37.21 | 94.92 | — | — | |
| MaxLogitNetwork=DenseNet2023.03 | 89.57 | 8.38 | 48.96 | 94.78 | — | — | |
| RegMixupNetwork=ResNet1102023.03 | 89.4 | 7.88 | 50.91 | 95.1 | — | — | |
| EnergyNetwork=DenseNet2023.03 | 89.21 | 8.6 | 51.31 | 94.78 | — | — | |
| LogitNormNetwork=ResNet1102023.03 | 88.82 | 12.57 | 56.27 | 92.64 | — | — | |
| LogitNormNetwork=DenseNet2023.03 | 88.7 | 10.89 | 56.59 | 93.59 | — | — | |
| MixupNetwork=ResNet1102023.03 | 86.21 | 16.27 | 40.71 | 94.69 | — | — | |
| MaxLogitNetwork=ResNet1102023.03 | 85 | 14.93 | 53.01 | 94.3 | — | — | |
| EnergyNetwork=ResNet1102023.03 | 84.72 | 15.13 | 53.89 | 94.3 | — | — | |
| ODINNetwork=DenseNet2023.03 | 82.02 | 15.37 | 61.77 | 94.78 | — | — | |
| G-ShapleyN=50,000, d=7682026.05 | 81 | — | — | — | — | — | |
| ODINNetwork=ResNet1102023.03 | 79.45 | 20.82 | 59.32 | 94.3 | — | — | |
| ODINNetwork=WRNet2023.03 | 74.7 | 20.37 | 62.04 | 95.91 | — | — | |
| DAEDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 99.08 | |
| DropoutBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 98.86 | |
| EDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 98.74 | |
| F-EDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 99.1 | |
| I-EDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 98.72 | |
| MoDEXBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 99.18 | |
| R-EDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 98.98 | |
| Re-EDLBackbone=VGG-16, Uncertainty Type=aleatoric2026.05 | — | — | — | — | — | 98.81 |