Dynamic Alignment Mask CTC: Improved Mask-CTC with Aligned Cross Entropy

Conformer Architecture of Dynamic Alignment Mask CTC with Mask and Dynamic Rectification Methods

Abstract

Because of predicting all the target tokens in parallel, the non-autoregressive models greatly improve the decoding efficiency of speech recognition compared with traditional autoregressive models. In this work, we present dynamic alignment Mask CTC, introducing two methods{:} (1) Aligned Cross Entropy (AXE), finding the monotonic alignment that minimizes the cross-entropy loss through dynamic programming, (2) Dynamic Rectification, creating new training samples by replacing some masks with model predicted tokens. The AXE ignores the absolute position alignment between prediction and ground truth sentence and focuses on tokens matching in relative order. The dynamic rectification method makes the model capable of simulating the non-mask but possible wrong tokens, even if they have high confidence. Our experiments on WSJ dataset demonstrated that not only AXE loss but also the rectification method could improve the WER performance of Mask CTC.

Type
Publication
In 2023 IEEE International Conference on Acoustics, Speech and Signal Processing
Click the Cite button above to demo the feature to enable visitors to import publication metadata into their reference management software.
Haobin Tang
Haobin Tang
University of Science and Technology of China
Ning Cheng
Ning Cheng
Director
Jian Luo
Jian Luo
Researcher