Noisy Student Training is a semi-supervised training method which achieves 88.4% top-1 accuracy on ImageNet and surprising gains on robustness and adversarial benchmarks. Noisy Student Training is based on the self-training framework and trained with 4-simple steps:
- Train a classifier on labeled data (teacher).
- Infer labels on a much larger unlabeled dataset.
- Train a larger classifier on the combined set, adding noise (noisy student).
- Go to step 2, with student as teacher.