Misclassification Detection on MNIST (test)
0.97AUROCSoftmax Baseline
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Softmax BaselineArchitecture=3-layer, 256 neuron-wide, fully-connected network, Nonlinearity=GELU, Optimizer=Adam2016.10 | 0.97 | 0.5 | 1 | 0.98 | 0.48 | 0.017 | 0.86 | |
| MC-Dropout2021.06 | 0.878 | — | — | — | — | — | — | |
| MC-Dropout2021.06 | 0.878 | — | — | — | — | — | — | |
| MC-Dropoutscore_function=Eq. 62021.06 | 0.878 | — | — | — | — | — | — | |
| ReDecNN2021.06 | 0.869 | — | — | — | — | — | — | |
| ReDecNN2021.06 | 0.869 | — | — | — | — | — | — | |
| ReDecNNscore_function=Eq. 62021.06 | 0.869 | — | — | — | — | — | — | |
| BayesNN2021.06 | 0.537 | — | — | — | — | — | — | |
| BayesNN2021.06 | 0.537 | — | — | — | — | — | — | |
| BayesNNscore_function=Eq. 62021.06 | 0.537 | — | — | — | — | — | — | |
| Ensemble2021.06 | 0.531 | — | — | — | — | — | — | |
| Ensemble2021.06 | 0.531 | — | — | — | — | — | — | |
| Ensemblescore_function=Eq. 62021.06 | 0.531 | — | — | — | — | — | — |