Image Classification on MNIST-M unbalanced + flexible + adversarial setting (test)
97.35Random Baseline AccuracyStableNet
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| StableNetsetting=unbalanced + flexible + adversarial2021.04 | 97.35 | 95.33 | 93.49 | 91.24 | 87.04 | 75.69 | 75.46 | 87.94 | |
| JiGensetting=unbalanced + flexible + adversarial2021.04 | 97.18 | 94.97 | 92.99 | 90.64 | 78.97 | 68.79 | 69.34 | 84.7 | |
| RSCsetting=unbalanced + flexible + adversarial2021.04 | 96.94 | 93.12 | 89.44 | 85.78 | 81.68 | 69.15 | 65.12 | 83.08 | |
| CNNssetting=unbalanced + flexible + adversarial2021.04 | 96.93 | 93.76 | 91.93 | 88.13 | 81.48 | 68.43 | 66.11 | 83.82 | |
| DG-MMLDsetting=unbalanced + flexible + adversarial2021.04 | 96.89 | 94.61 | 92.59 | 89.72 | 88.44 | 69.13 | 71.39 | 86.11 | |
| M-ADAsetting=unbalanced + flexible + adversarial2021.04 | 95.92 | 94.45 | 92.29 | 88.87 | 85.89 | 70.32 | 67.08 | 84.97 |