Pushing the performances of ASR models on English and Spanish accents

Speech to text models tend to be trained and evaluated against a single target accent. This is especially true for English for which native speakers from the United States became the main benchmark. In this work, we are going to show how two simple methods: pre-trained embeddings and auxiliary class...

Full description

Saved in:
Bibliographic Details
Main Authors Chitkara, Pooja, Riviere, Morgane, Copet, Jade, Zhang, Frank, Saraf, Yatharth
Format Journal Article
LanguageEnglish
Published 22.12.2022
Subjects
Computer Science - Computation and Language
Computer Science - Sound
Online AccessGet full text

Cover

More Information
Summary:Speech to text models tend to be trained and evaluated against a single target accent. This is especially true for English for which native speakers from the United States became the main benchmark. In this work, we are going to show how two simple methods: pre-trained embeddings and auxiliary classification losses can improve the performance of ASR systems. We are looking for upgrades as universal as possible and therefore we will explore their impact on several models architectures and several languages.
DOI:10.48550/arxiv.2212.12048