Enhancing Code-Switching ASR Leveraging Non-Peaky CTC Loss and Deep Language Posterior Injection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Tzu-Ting, Wang, Hsin-Wei, Wang, Yi-Cheng, Chen, Berlin
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909425213636608
author Yang, Tzu-Ting
Wang, Hsin-Wei
Wang, Yi-Cheng
Chen, Berlin
author_facet Yang, Tzu-Ting
Wang, Hsin-Wei
Wang, Yi-Cheng
Chen, Berlin
contents Code-switching-where multilingual speakers alternately switch between languages during conversations-still poses significant challenges to end-to-end (E2E) automatic speech recognition (ASR) systems due to phenomena of both acoustic and semantic confusion. This issue arises because ASR systems struggle to handle the rapid alternation of languages effectively, which often leads to significant performance degradation. Our main contributions are at least threefold: First, we incorporate language identification (LID) information into several intermediate layers of the encoder, aiming to enrich output embeddings with more detailed language information. Secondly, through the novel application of language boundary alignment loss, the subsequent ASR modules are enabled to more effectively utilize the knowledge of internal language posteriors. Third, we explore the feasibility of using language posteriors to facilitate deep interaction between shared encoder and language-specific encoders. Through comprehensive experiments on the SEAME corpus, we have verified that our proposed method outperforms the prior-art method, disentangle based mixture-of-experts (D-MoE), further enhancing the acuity of the encoder to languages.
format Preprint
id arxiv_https___arxiv_org_abs_2412_08651
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing Code-Switching ASR Leveraging Non-Peaky CTC Loss and Deep Language Posterior Injection
Yang, Tzu-Ting
Wang, Hsin-Wei
Wang, Yi-Cheng
Chen, Berlin
Audio and Speech Processing
Computation and Language
Machine Learning
Sound
Code-switching-where multilingual speakers alternately switch between languages during conversations-still poses significant challenges to end-to-end (E2E) automatic speech recognition (ASR) systems due to phenomena of both acoustic and semantic confusion. This issue arises because ASR systems struggle to handle the rapid alternation of languages effectively, which often leads to significant performance degradation. Our main contributions are at least threefold: First, we incorporate language identification (LID) information into several intermediate layers of the encoder, aiming to enrich output embeddings with more detailed language information. Secondly, through the novel application of language boundary alignment loss, the subsequent ASR modules are enabled to more effectively utilize the knowledge of internal language posteriors. Third, we explore the feasibility of using language posteriors to facilitate deep interaction between shared encoder and language-specific encoders. Through comprehensive experiments on the SEAME corpus, we have verified that our proposed method outperforms the prior-art method, disentangle based mixture-of-experts (D-MoE), further enhancing the acuity of the encoder to languages.
title Enhancing Code-Switching ASR Leveraging Non-Peaky CTC Loss and Deep Language Posterior Injection
topic Audio and Speech Processing
Computation and Language
Machine Learning
Sound
url https://arxiv.org/abs/2412.08651