%0 Conference Proceedings %T Self-Training with Differentiable Teacher %A Zuo, Simiao %A Yu, Yue %A Liang, Chen %A Jiang, Haoming %A Er, Siawpeng %A Zhang, Chao %A Zhao, Tuo %A Zha, Hongyuan %Y Carpuat, Marine %Y de Marneffe, Marie-Catherine %Y Meza Ruiz, Ivan Vladimir %S Findings of the Association for Computational Linguistics: NAACL 2022 %D 2022 %8 July %I Association for Computational Linguistics %C Seattle, United States %F zuo-etal-2022-self %X Self-training achieves enormous success in various semi-supervised and weakly-supervised learning tasks. The method can be interpreted as a teacher-student framework, where the teacher generates pseudo-labels, and the student makes predictions. The two models are updated alternatingly. However, such a straightforward alternating update rule leads to training instability. This is because a small change in the teacher may result in a significant change in the student. To address this issue, we propose DRIFT, short for differentiable self-training, that treats teacher-student as a Stackelberg game. In this game, a leader is always in a more advantageous position than a follower. In self-training, the student contributes to the prediction performance, and the teacher controls the training process by generating pseudo-labels. Therefore, we treat the student as the leader and the teacher as the follower. The leader procures its advantage by acknowledging the follower’s strategy, which involves differentiable pseudo-labels and differentiable sample weights. Consequently, the leader-follower interaction can be effectively captured via Stackelberg gradient, obtained by differentiating the follower’s strategy. Experimental results on semi- and weakly-supervised classification and named entity recognition tasks show that our model outperforms existing approaches by large margins. %R 10.18653/v1/2022.findings-naacl.70 %U https://aclanthology.org/2022.findings-naacl.70 %U https://doi.org/10.18653/v1/2022.findings-naacl.70 %P 933-949