Single-Source Domain Generalization on Digits (test)
87.76Acc (MNIST-M)RandConv
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RandConvNumber of convolutional-based augmentation layers=12023.07 | 87.76 | 98.85 | 57.62 | 83.36 | 62.88 | 72.88 | |
| ESDA2023.07 | 81.6 | 99.3 | 48.9 | 84 | 62.2 | 69.12 | |
| ABANumber of convolutional-based augmentation layers=32023.07 | 80.94 | 98.73 | 55.88 | 96.34 | 73.09 | 76.56 | |
| ABANumber of convolutional-based augmentation layers=52023.07 | 80.54 | 98.78 | 52.45 | 95.81 | 70.25 | 74.76 | |
| ABA+AugmixNumber of convolutional-based augmentation layers=5, Augmentation module=Augmix2023.07 | 80.24 | 98.66 | 56.43 | 96.14 | 70.91 | 75.93 | |
| ABA+RandConvNumber of convolutional-based augmentation layers=3, Augmentation module=RandConv2023.07 | 80.05 | 98.67 | 56.87 | 96.55 | 73.4 | 76.72 | |
| ABA+RandConvNumber of convolutional-based augmentation layers=5, Augmentation module=RandConv2023.07 | 79.69 | 98.76 | 54.09 | 96.42 | 71.55 | 75.44 | |
| ABANumber of convolutional-based augmentation layers=12023.07 | 78.81 | 98.82 | 51.88 | 96.22 | 71.25 | 74.57 | |
| ABA+RandConvNumber of convolutional-based augmentation layers=1, Augmentation module=RandConv2023.07 | 78.62 | 98.78 | 52.04 | 96.16 | 71.23 | 74.51 | |
| ALT+RandConvNumber of convolutional-based augmentation layers=5, Augmentation module=RandConv2023.07 | 76.9 | 98.46 | 53.78 | 95.4 | 69.4 | 73.87 | |
| ALT+AugmixNumber of convolutional-based augmentation layers=5, Augmentation module=Augmix2023.07 | 75.98 | 98.55 | 55.01 | 96.17 | 68.93 | 74.38 | |
| ALT+RandConvNumber of convolutional-based augmentation layers=1, Augmentation module=RandConv2023.07 | 75.77 | 98.54 | 49.9 | 95.64 | 68.61 | 72.47 | |
| ALTNumber of convolutional-based augmentation layers=52023.07 | 74.28 | 98.46 | 52.25 | 94.99 | 68.44 | 72.49 | |
| ALTNumber of convolutional-based augmentation layers=12023.07 | 72.8 | 98.41 | 47.07 | 94.79 | 66.27 | 70.23 | |
| AdvBNN2023.07 | 71.79 | 98.23 | 44.85 | 46.05 | 44.99 | 51.92 | |
| M-ADA2023.07 | 67.94 | 99.3 | 42.55 | 78.53 | 48.95 | 59.49 | |
| ADA2023.07 | 60.41 | — | 35.51 | 77.26 | 45.32 | 54.62 | |
| ERM2023.07 | 58.87 | 98.4 | 33.41 | 79.27 | 42.43 | 53.5 | |
| Augmix2023.07 | 53.36 | 98.53 | 25.96 | 96.12 | 42.9 | 54.59 |