Hierarchical Conditional End-to-End ASR with CTC and Multi-Granular Subword Units

In end-to-end automatic speech recognition (ASR), a model is expected to implicitly learn representations suitable for recognizing a word-level sequence. However, the huge abstraction gap between input acoustic signals and output linguistic tokens makes it challenging for a model to learn the repres...

Full description

Saved in:
Bibliographic Details
Published inICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 7797 - 7801
Main Authors Higuchi, Yosuke, Karube, Keita, Ogawa, Tetsuji, Kobayashi, Tetsunori
Format Conference Proceeding
LanguageEnglish
Published IEEE 23.05.2022
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:In end-to-end automatic speech recognition (ASR), a model is expected to implicitly learn representations suitable for recognizing a word-level sequence. However, the huge abstraction gap between input acoustic signals and output linguistic tokens makes it challenging for a model to learn the representations. In this work, to promote the word-level representation learning in end-to-end ASR, we propose a hierarchical conditional model that is based on connectionist temporal classification (CTC). Our model is trained by auxiliary CTC losses applied to intermediate layers, where the vocabulary size of each target subword sequence is gradually increased as the layer becomes close to the word-level output. Here, we make each level of sequence prediction explicitly conditioned on the previous sequences predicted at lower levels. With the proposed approach, we expect the proposed model to learn the word-level representations effectively by exploiting a hierarchy of linguistic structures. Experimental results on LibriSpeech-{100h, 960h} and TEDLIUM2 demonstrate that the proposed model improves over a standard CTCbased model and other competitive models from prior work. We further analyze the results to confirm the effectiveness of the intended representation learning with our model.
ISSN:2379-190X
DOI:10.1109/ICASSP43922.2022.9746580