DMT: Dynamic Mutual Training for Semi-Supervised Learning
About
Recent semi-supervised learning methods use pseudo supervision as core idea, especially self-training methods that generate pseudo labels. However, pseudo labels are unreliable. Self-training methods usually rely on single model prediction confidence to filter low-confidence pseudo labels, thus remaining high-confidence errors and wasting many low-confidence correct labels. In this paper, we point out it is difficult for a model to counter its own errors. Instead, leveraging inter-model disagreement between different models is a key to locate pseudo label errors. With this new viewpoint, we propose mutual training between two different models by a dynamically re-weighted loss function, called Dynamic Mutual Training (DMT). We quantify inter-model disagreement by comparing predictions from two different models to dynamically re-weight loss in training, where a larger disagreement indicates a possible error and corresponds to a lower loss value. Extensive experiments show that DMT achieves state-of-the-art performance in both image classification and semantic segmentation. Our codes are released at https://github.com/voldemortX/DST-CBC .
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Semantic segmentation | PASCAL VOC 2012 (val) | Mean IoU74.75 | 2040 | |
| Semantic segmentation | GTA5 → Cityscapes (val) | mIoU63.2 | 533 | |
| Semantic segmentation | PASCAL VOC (val) | mIoU74.8 | 338 | |
| Semantic segmentation | Cityscapes (val) | mIoU66.9 | 332 | |
| Semantic segmentation | Cityscapes (val) | mIoU68.2 | 133 | |
| Medical Image Segmentation | Kvasir-SEG (test) | mIoU79.32 | 78 | |
| Semantic segmentation | SYNTHIA-to-Cityscapes 16 categories (val) | mIoU (Overall)59.7 | 74 | |
| Semantic segmentation | CITYSCAPES 1/8 labeled samples 372 labels (val) | mIoU63 | 65 | |
| Image Classification | CIFAR-10 4,000 labels (test) | -- | 57 | |
| Semantic segmentation | Cityscapes fine (val) | mIoU68.16 | 42 |