Improving Language Identification of Accented Speech

Language identification from speech is a common preprocessing step in many spoken language processing systems. In recent years, this field has seen fast progress, mostly due to the use of self-supervised models pretrained on multilingual data and the use of large training corpora. This paper shows t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kukk, Kunnar, Alumäe, Tanel
Format: Artikel
Sprache:eng
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Kukk, Kunnar
Alumäe, Tanel
description Language identification from speech is a common preprocessing step in many spoken language processing systems. In recent years, this field has seen fast progress, mostly due to the use of self-supervised models pretrained on multilingual data and the use of large training corpora. This paper shows that for speech with a non-native or regional accent, the accuracy of spoken language identification systems drops dramatically, and that the accuracy of identifying the language is inversely correlated with the strength of the accent. We also show that using the output of a lexicon-free speech recognition system of the particular language helps to improve language identification performance on accented speech by a large margin, without sacrificing accuracy on native speech. We obtain relative error rate reductions ranging from to 35 to 63% over the state-of-the-art model across several non-native speech datasets.
doi_str_mv 10.48550/arxiv.2203.16972
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2203_16972</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2203_16972</sourcerecordid><originalsourceid>FETCH-LOGICAL-a672-1af081c2e0e8562583701dfe463f89ef1c56705ad1b790aa8c40be2e0446f3163</originalsourceid><addsrcrecordid>eNotzrtqw0AUBNBtUgQnH5DK-wNS7r5XpTB5CAQurF5cr-7KC9EDxTH238dxXA0Mw3AYexGQa28MvOJyTqdcSlC5sIWTj0xXw7xMpzT2vMax_8GeeNXReEwxBTymaeRT5GUI14o6vpuJwuGJPUT8-qbne65Y8_7WbD6zevtRbco6Q-tkJjCCF0ESkDdWGq8ciC6Stir6gqIIxjow2Im9KwDRBw17us61tlEJq1Zs_X97Y7fzkgZcLu0fv73x1S8yhT6_</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Improving Language Identification of Accented Speech</title><source>arXiv.org</source><creator>Kukk, Kunnar ; Alumäe, Tanel</creator><creatorcontrib>Kukk, Kunnar ; Alumäe, Tanel</creatorcontrib><description>Language identification from speech is a common preprocessing step in many spoken language processing systems. In recent years, this field has seen fast progress, mostly due to the use of self-supervised models pretrained on multilingual data and the use of large training corpora. This paper shows that for speech with a non-native or regional accent, the accuracy of spoken language identification systems drops dramatically, and that the accuracy of identifying the language is inversely correlated with the strength of the accent. We also show that using the output of a lexicon-free speech recognition system of the particular language helps to improve language identification performance on accented speech by a large margin, without sacrificing accuracy on native speech. We obtain relative error rate reductions ranging from to 35 to 63% over the state-of-the-art model across several non-native speech datasets.</description><identifier>DOI: 10.48550/arxiv.2203.16972</identifier><language>eng</language><creationdate>2022-03</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2203.16972$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2203.16972$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Kukk, Kunnar</creatorcontrib><creatorcontrib>Alumäe, Tanel</creatorcontrib><title>Improving Language Identification of Accented Speech</title><description>Language identification from speech is a common preprocessing step in many spoken language processing systems. In recent years, this field has seen fast progress, mostly due to the use of self-supervised models pretrained on multilingual data and the use of large training corpora. This paper shows that for speech with a non-native or regional accent, the accuracy of spoken language identification systems drops dramatically, and that the accuracy of identifying the language is inversely correlated with the strength of the accent. We also show that using the output of a lexicon-free speech recognition system of the particular language helps to improve language identification performance on accented speech by a large margin, without sacrificing accuracy on native speech. We obtain relative error rate reductions ranging from to 35 to 63% over the state-of-the-art model across several non-native speech datasets.</description><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzrtqw0AUBNBtUgQnH5DK-wNS7r5XpTB5CAQurF5cr-7KC9EDxTH238dxXA0Mw3AYexGQa28MvOJyTqdcSlC5sIWTj0xXw7xMpzT2vMax_8GeeNXReEwxBTymaeRT5GUI14o6vpuJwuGJPUT8-qbne65Y8_7WbD6zevtRbco6Q-tkJjCCF0ESkDdWGq8ciC6Stir6gqIIxjow2Im9KwDRBw17us61tlEJq1Zs_X97Y7fzkgZcLu0fv73x1S8yhT6_</recordid><startdate>20220331</startdate><enddate>20220331</enddate><creator>Kukk, Kunnar</creator><creator>Alumäe, Tanel</creator><scope>GOX</scope></search><sort><creationdate>20220331</creationdate><title>Improving Language Identification of Accented Speech</title><author>Kukk, Kunnar ; Alumäe, Tanel</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a672-1af081c2e0e8562583701dfe463f89ef1c56705ad1b790aa8c40be2e0446f3163</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><toplevel>online_resources</toplevel><creatorcontrib>Kukk, Kunnar</creatorcontrib><creatorcontrib>Alumäe, Tanel</creatorcontrib><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Kukk, Kunnar</au><au>Alumäe, Tanel</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Improving Language Identification of Accented Speech</atitle><date>2022-03-31</date><risdate>2022</risdate><abstract>Language identification from speech is a common preprocessing step in many spoken language processing systems. In recent years, this field has seen fast progress, mostly due to the use of self-supervised models pretrained on multilingual data and the use of large training corpora. This paper shows that for speech with a non-native or regional accent, the accuracy of spoken language identification systems drops dramatically, and that the accuracy of identifying the language is inversely correlated with the strength of the accent. We also show that using the output of a lexicon-free speech recognition system of the particular language helps to improve language identification performance on accented speech by a large margin, without sacrificing accuracy on native speech. We obtain relative error rate reductions ranging from to 35 to 63% over the state-of-the-art model across several non-native speech datasets.</abstract><doi>10.48550/arxiv.2203.16972</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2203.16972
ispartof
issn
language eng
recordid cdi_arxiv_primary_2203_16972
source arXiv.org
title Improving Language Identification of Accented Speech
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-14T02%3A43%3A54IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Improving%20Language%20Identification%20of%20Accented%20Speech&rft.au=Kukk,%20Kunnar&rft.date=2022-03-31&rft_id=info:doi/10.48550/arxiv.2203.16972&rft_dat=%3Carxiv_GOX%3E2203_16972%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true