Many semi- and weakly-supervised approaches have been investigated for overcoming the labeling cost of building high quality speech recognition systems. On the challenging task of transcribing social media videos in low-resource conditions, we conduct a large scale systematic comparison between two self-labeling methods on one hand, and weakly-supervised pretraining using contextual metadata on the other. We investigate distillation methods at the frame level and the sequence level for hybrid, encoder-only CTC-based, and encoder-decoder speech recognition systems on Dutch and Romanian languages using 27,000 and 58,000 hours of unlabeled audio respectively. Although all approaches improved upon their respective baseline WERs by more than 8%, sequence-level distillation for encoder-decoder models provided the largest relative WER reduction of 20% compared to the strongest data-augmented supervised baseline.
@article{arxiv.2005.07850,
title = {Large scale weakly and semi-supervised learning for low-resource video ASR},
author = {Kritika Singh and Vimal Manohar and Alex Xiao and Sergey Edunov and Ross Girshick and Vitaliy Liptchinsky and Christian Fuegen and Yatharth Saraf and Geoffrey Zweig and Abdelrahman Mohamed},
journal= {arXiv preprint arXiv:2005.07850},
year = {2020}
}