@inproceedings{jimerson-etal-2023-unhelpful,
title = "An (unhelpful) guide to selecting the best {ASR} architecture for your under-resourced language",
author = "Jimerson, Robert and
Liu, Zoey and
Prud{'}hommeaux, Emily",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-short.87",
doi = "10.18653/v1/2023.acl-short.87",
pages = "1008--1016",
abstract = "Advances in deep neural models for automatic speech recognition (ASR) have yielded dramatic improvements in ASR quality for resource-rich languages, with English ASR now achieving word error rates comparable to that of human transcribers. The vast majority of the world{'}s languages, however, lack the quantity of data necessary to approach this level of accuracy. In this paper we use four of the most popular ASR toolkits to train ASR models for eleven languages with limited ASR training resources: eleven widely spoken languages of Africa, Asia, and South America, one endangered language of Central America, and three critically endangered languages of North America. We find that no single architecture consistently outperforms any other. These differences in performance so far do not appear to be related to any particular feature of the datasets or characteristics of the languages. These findings have important implications for future research in ASR for under-resourced languages. ASR systems for languages with abundant existing media and available speakers may derive the most benefit simply by collecting large amounts of additional acoustic and textual training data. Communities using ASR to support endangered language documentation efforts, who cannot easily collect more data, might instead focus on exploring multiple architectures and hyperparameterizations to optimize performance within the constraints of their available data and resources.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jimerson-etal-2023-unhelpful">
<titleInfo>
<title>An (unhelpful) guide to selecting the best ASR architecture for your under-resourced language</title>
</titleInfo>
<name type="personal">
<namePart type="given">Robert</namePart>
<namePart type="family">Jimerson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zoey</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Emily</namePart>
<namePart type="family">Prud’hommeaux</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Advances in deep neural models for automatic speech recognition (ASR) have yielded dramatic improvements in ASR quality for resource-rich languages, with English ASR now achieving word error rates comparable to that of human transcribers. The vast majority of the world’s languages, however, lack the quantity of data necessary to approach this level of accuracy. In this paper we use four of the most popular ASR toolkits to train ASR models for eleven languages with limited ASR training resources: eleven widely spoken languages of Africa, Asia, and South America, one endangered language of Central America, and three critically endangered languages of North America. We find that no single architecture consistently outperforms any other. These differences in performance so far do not appear to be related to any particular feature of the datasets or characteristics of the languages. These findings have important implications for future research in ASR for under-resourced languages. ASR systems for languages with abundant existing media and available speakers may derive the most benefit simply by collecting large amounts of additional acoustic and textual training data. Communities using ASR to support endangered language documentation efforts, who cannot easily collect more data, might instead focus on exploring multiple architectures and hyperparameterizations to optimize performance within the constraints of their available data and resources.</abstract>
<identifier type="citekey">jimerson-etal-2023-unhelpful</identifier>
<identifier type="doi">10.18653/v1/2023.acl-short.87</identifier>
<location>
<url>https://aclanthology.org/2023.acl-short.87</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>1008</start>
<end>1016</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T An (unhelpful) guide to selecting the best ASR architecture for your under-resourced language
%A Jimerson, Robert
%A Liu, Zoey
%A Prud’hommeaux, Emily
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F jimerson-etal-2023-unhelpful
%X Advances in deep neural models for automatic speech recognition (ASR) have yielded dramatic improvements in ASR quality for resource-rich languages, with English ASR now achieving word error rates comparable to that of human transcribers. The vast majority of the world’s languages, however, lack the quantity of data necessary to approach this level of accuracy. In this paper we use four of the most popular ASR toolkits to train ASR models for eleven languages with limited ASR training resources: eleven widely spoken languages of Africa, Asia, and South America, one endangered language of Central America, and three critically endangered languages of North America. We find that no single architecture consistently outperforms any other. These differences in performance so far do not appear to be related to any particular feature of the datasets or characteristics of the languages. These findings have important implications for future research in ASR for under-resourced languages. ASR systems for languages with abundant existing media and available speakers may derive the most benefit simply by collecting large amounts of additional acoustic and textual training data. Communities using ASR to support endangered language documentation efforts, who cannot easily collect more data, might instead focus on exploring multiple architectures and hyperparameterizations to optimize performance within the constraints of their available data and resources.
%R 10.18653/v1/2023.acl-short.87
%U https://aclanthology.org/2023.acl-short.87
%U https://doi.org/10.18653/v1/2023.acl-short.87
%P 1008-1016
Markdown (Informal)
[An (unhelpful) guide to selecting the best ASR architecture for your under-resourced language](https://aclanthology.org/2023.acl-short.87) (Jimerson et al., ACL 2023)
ACL