Image Classification and OOD Detection on CIFAR-10 (ID) vs CIFAR-100 (OOD) (test)
98.64ID AccuracyC-EDL
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| C-EDLUncertainty Strategy=Monte Carlo2025.06 | 98.64 | 45.02 | 11.44 | 7.2 | |
| C-EDLUncertainty Strategy=Metamorphic Transformation2025.06 | 98.22 | 53.18 | 9.61 | 3.17 | |
| EDL++Uncertainty Strategy=Monte Carlo2025.06 | 97.87 | 54.5 | 19.5 | 16.32 | |
| EDL++Uncertainty Strategy=Metamorphic Transformation2025.06 | 97.51 | 59.47 | 13.34 | 5.46 | |
| R-EDL2025.06 | 97.4 | 60.21 | 14.46 | 9.4 | |
| H-EDL2025.06 | 96.76 | 62.25 | 16.12 | 12.73 | |
| S-EDL2025.06 | 96.69 | 63.13 | 17.48 | 7.73 | |
| I-EDL2025.06 | 96.67 | 63.84 | 17.73 | 14.1 | |
| EDL2025.06 | 96.66 | 63.33 | 17.24 | 14.02 | |
| Posterior Network2025.06 | 96.17 | 62.91 | 17.73 | 12.3 | |
| DA-EDL2025.06 | 89.28 | 66.98 | 33.85 | 31.88 |