Out-of-distribution detection on CIFAR-10 Gaussian
100AUROCMulti-Task Curriculum Framework
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Multi-Task Curriculum Framework2020.07 | 100 | — | — | — | — | — | — | — | |
| ODINBackbone=Wide ResNet 28x102021.05 | 100 | — | — | — | — | — | — | 0 | |
| Deep Abstaining ClassifierBackbone=ResNet 342021.05 | 99.98 | — | — | — | — | — | — | 4 | |
| Deep Abstaining ClassifierBackbone=Wide ResNet 28x102021.05 | 99.95 | — | — | — | — | — | — | 13 | |
| Deep Abstaining ClassifierBackbone=Wide ResNet 40x22021.05 | 99.93 | — | — | — | — | — | — | 13 | |
| Outlier ExposureBackbone=Wide ResNet 40x22021.05 | 99.6 | — | — | — | — | — | — | 70 | |
| Ensemble of Leave-out ClassifiersBackbone=Wide ResNet 28x102021.05 | 99.58 | — | — | — | — | — | — | 0 | |
| Softmax BaselineArchitecture=40-4 Wide ResNet, Nonlinearity=GELU, Optimizer=SGD with restarts, Training Epochs=502016.10 | 97 | 50 | 98 | 49 | 95 | 51 | 77 | — | |
| OpenMaxBackbone=ResNet 342021.05 | 84.74 | — | — | — | — | — | — | 4,058 | |
| ODIN2020.07 | 37.49 | — | — | — | — | — | — | — | |
| Hendrycks & Gimpel2020.07 | 32.41 | — | — | — | — | — | — | — |