How unlabeled data improve generalization in self-training? A one-hidden-layer theoretical analysis.

Shuai Zhang, Meng Wang, Sijia Liu, Pin-Yu Chen, Jinjun Xiong

Browse the full ICLR paper archive.