Novelty Detection on CIFAR-10 vs CIFAR-100
0.9AUROCFITYMI
Evaluation Results
| Method | Links | |
|---|---|---|
| FITYMITraining=Pre-trained2022.05 | 0.9 | |
| TransformalyTraining=Pre-trained2022.05 | 0.823 | |
| MSADTraining=Pre-trained2022.05 | 0.795 | |
| PANDATraining=Pre-trained2022.05 | 0.768 | |
| CSITraining=From Scratch2022.05 | 0.761 | |
| MHRotTraining=From Scratch, Require Extra Outlier Exposure Datasets=true2022.05 | 0.631 | |
| GTTraining=From Scratch, Require Extra Outlier Exposure Datasets=true2022.05 | 0.626 | |
| ADIBTraining=Pre-trained2022.05 | 0.607 | |
| DN2Training=Pre-trained2022.05 | 0.585 | |
| DeepSADTraining=From Scratch, Require Extra Outlier Exposure Datasets=true2022.05 | 0.5 | |
| One-class-OpenGANTraining=Pre-trained2022.05 | 0.5 |