Spoken Term Discovery for Language Documentation using Translations

Antonios Anastasopoulos; Sameer Bansal; Sharon Goldwater; Adam Lopez; David Chiang

doi:10.18653/v1/W17-4607

Spoken Term Discovery for Language Documentation using Translations

Antonios Anastasopoulos, Sameer Bansal, Sharon Goldwater, Adam Lopez, David Chiang

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

Vast amounts of speech data collected for language documentation and research remain untranscribed and unsearchable, but often a small amount of speech may have text translations available. We present a method for partially labeling additional speech with translations in this scenario. We modify an unsupervised speech-totranslation alignment model and obtain prototype speech segments that match the translation words, which are in turn used to discover terms in the unlabelled data. We evaluate our method on a Spanish-English speech translation corpus and on two corpora of endangered languages, Arapaho and Ainu, demonstrating its appropriateness and applicability in an actual very-low-resource scenario.

Original language	English
Title of host publication	Proceedings of the First Workshop on Speech-Centric Natural Language Processing
Publisher	Association for Computational Linguistics
Pages	53-58
Number of pages	6
ISBN (Print)	978-1-945626-92-0
DOIs	https://doi.org/10.18653/v1/W17-4607
Publication status	Published - 11 Sept 2017
Event	First Workshop on Speech-Centric Natural Language Processing - Copenhagen, Denmark Duration: 7 Sept 2017 → 7 Sept 2017 http://speechnlp.github.io/2017/

Conference

Conference	First Workshop on Speech-Centric Natural Language Processing
Abbreviated title	SCNLP 2017
Country/Territory	Denmark
City	Copenhagen
Period	7/09/17 → 7/09/17
Internet address	http://speechnlp.github.io/2017/

Access to Document

10.18653/v1/W17-4607Licence: Creative Commons: Attribution (CC-BY)

Keyword_Spotting_for_Endangered_Languages-1Accepted author manuscript, 87.4 KB

Cite this

@inproceedings{c22311dac2bb49c885f28e8dce954745,

title = "Spoken Term Discovery for Language Documentation using Translations",

abstract = "Vast amounts of speech data collected for language documentation and research remain untranscribed and unsearchable, but often a small amount of speech may have text translations available. We present a method for partially labeling additional speech with translations in this scenario. We modify an unsupervised speech-totranslation alignment model and obtain prototype speech segments that match the translation words, which are in turn used to discover terms in the unlabelled data. We evaluate our method on a Spanish-English speech translation corpus and on two corpora of endangered languages, Arapaho and Ainu, demonstrating its appropriateness and applicability in an actual very-low-resource scenario.",

author = "Antonios Anastasopoulos and Sameer Bansal and Sharon Goldwater and Adam Lopez and David Chiang",

year = "2017",

month = sep,

day = "11",

doi = "10.18653/v1/W17-4607",

language = "English",

isbn = "978-1-945626-92-0",

pages = "53--58",

booktitle = "Proceedings of the First Workshop on Speech-Centric Natural Language Processing",

publisher = "Association for Computational Linguistics",

note = "First Workshop on Speech-Centric Natural Language Processing, SCNLP 2017 ; Conference date: 07-09-2017 Through 07-09-2017",

url = "http://speechnlp.github.io/2017/",

}

Anastasopoulos, A, Bansal, S, Goldwater, S , Lopez, A & Chiang, D 2017, Spoken Term Discovery for Language Documentation using Translations. in Proceedings of the First Workshop on Speech-Centric Natural Language Processing. Association for Computational Linguistics, pp. 53-58, First Workshop on Speech-Centric Natural Language Processing, Copenhagen, Denmark, 7/09/17. https://doi.org/10.18653/v1/W17-4607

TY - GEN

T1 - Spoken Term Discovery for Language Documentation using Translations

AU - Anastasopoulos, Antonios

AU - Bansal, Sameer

AU - Goldwater, Sharon

AU - Lopez, Adam

AU - Chiang, David

PY - 2017/9/11

Y1 - 2017/9/11

N2 - Vast amounts of speech data collected for language documentation and research remain untranscribed and unsearchable, but often a small amount of speech may have text translations available. We present a method for partially labeling additional speech with translations in this scenario. We modify an unsupervised speech-totranslation alignment model and obtain prototype speech segments that match the translation words, which are in turn used to discover terms in the unlabelled data. We evaluate our method on a Spanish-English speech translation corpus and on two corpora of endangered languages, Arapaho and Ainu, demonstrating its appropriateness and applicability in an actual very-low-resource scenario.

AB - Vast amounts of speech data collected for language documentation and research remain untranscribed and unsearchable, but often a small amount of speech may have text translations available. We present a method for partially labeling additional speech with translations in this scenario. We modify an unsupervised speech-totranslation alignment model and obtain prototype speech segments that match the translation words, which are in turn used to discover terms in the unlabelled data. We evaluate our method on a Spanish-English speech translation corpus and on two corpora of endangered languages, Arapaho and Ainu, demonstrating its appropriateness and applicability in an actual very-low-resource scenario.

U2 - 10.18653/v1/W17-4607

DO - 10.18653/v1/W17-4607

M3 - Conference contribution

SN - 978-1-945626-92-0

SP - 53

EP - 58

BT - Proceedings of the First Workshop on Speech-Centric Natural Language Processing

PB - Association for Computational Linguistics

T2 - First Workshop on Speech-Centric Natural Language Processing

Y2 - 7 September 2017 through 7 September 2017

ER -

Spoken Term Discovery for Language Documentation using Translations

Abstract / Description of output

Conference

Access to Document

Fingerprint

Cite this