@inproceedings{tonja-etal-2023-parallel,
title = "Parallel Corpus for Indigenous Language Translation: {S}panish-Mazatec and {S}panish-{M}ixtec",
author = "Tonja, Atnafu Lambebo and
Maldonado-sifuentes, Christian and
Mendoza Castillo, David Alejandro and
Kolesnikova, Olga and
Castro-S{\'a}nchez, No{\'e} and
Sidorov, Grigori and
Gelbukh, Alexander",
editor = "Mager, Manuel and
Ebrahimi, Abteen and
Oncevay, Arturo and
Rice, Enora and
Rijhwani, Shruti and
Palmer, Alexis and
Kann, Katharina",
booktitle = "Proceedings of the Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.americasnlp-1.11/",
doi = "10.18653/v1/2023.americasnlp-1.11",
pages = "94--102",
abstract = "In this paper, we present a parallel Spanish- Mazatec and Spanish-Mixtec corpus for machine translation (MT) tasks, where Mazatec and Mixtec are two indigenous Mexican languages. We evaluated the usability of the collected corpus using three different approaches: transformer, transfer learning, and fine-tuning pre-trained multilingual MT models. Fine-tuning the Facebook m2m100-48 model outperformed the other approaches, with BLEU scores of 12.09 and 22.25 for Mazatec-Spanish and Spanish-Mazatec translations, respectively, and 16.75 and 22.15 for Mixtec-Spanish and Spanish-Mixtec translations, respectively. The results indicate that translation performance is influenced by the dataset size (9,799 sentences in Mazatec and 13,235 sentences in Mixtec) and is more effective when indigenous languages are used as target languages. The findings emphasize the importance of creating parallel corpora for indigenous languages and fine-tuning models for low-resource translation tasks. Future research will investigate zero-shot and few-shot learning approaches to further improve translation performance in low-resource settings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tonja-etal-2023-parallel">
<titleInfo>
<title>Parallel Corpus for Indigenous Language Translation: Spanish-Mazatec and Spanish-Mixtec</title>
</titleInfo>
<name type="personal">
<namePart type="given">Atnafu</namePart>
<namePart type="given">Lambebo</namePart>
<namePart type="family">Tonja</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christian</namePart>
<namePart type="family">Maldonado-sifuentes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="given">Alejandro</namePart>
<namePart type="family">Mendoza Castillo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Olga</namePart>
<namePart type="family">Kolesnikova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Noé</namePart>
<namePart type="family">Castro-Sánchez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Grigori</namePart>
<namePart type="family">Sidorov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alexander</namePart>
<namePart type="family">Gelbukh</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 Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Manuel</namePart>
<namePart type="family">Mager</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abteen</namePart>
<namePart type="family">Ebrahimi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arturo</namePart>
<namePart type="family">Oncevay</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Enora</namePart>
<namePart type="family">Rice</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shruti</namePart>
<namePart type="family">Rijhwani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alexis</namePart>
<namePart type="family">Palmer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Katharina</namePart>
<namePart type="family">Kann</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>In this paper, we present a parallel Spanish- Mazatec and Spanish-Mixtec corpus for machine translation (MT) tasks, where Mazatec and Mixtec are two indigenous Mexican languages. We evaluated the usability of the collected corpus using three different approaches: transformer, transfer learning, and fine-tuning pre-trained multilingual MT models. Fine-tuning the Facebook m2m100-48 model outperformed the other approaches, with BLEU scores of 12.09 and 22.25 for Mazatec-Spanish and Spanish-Mazatec translations, respectively, and 16.75 and 22.15 for Mixtec-Spanish and Spanish-Mixtec translations, respectively. The results indicate that translation performance is influenced by the dataset size (9,799 sentences in Mazatec and 13,235 sentences in Mixtec) and is more effective when indigenous languages are used as target languages. The findings emphasize the importance of creating parallel corpora for indigenous languages and fine-tuning models for low-resource translation tasks. Future research will investigate zero-shot and few-shot learning approaches to further improve translation performance in low-resource settings.</abstract>
<identifier type="citekey">tonja-etal-2023-parallel</identifier>
<identifier type="doi">10.18653/v1/2023.americasnlp-1.11</identifier>
<location>
<url>https://aclanthology.org/2023.americasnlp-1.11/</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>94</start>
<end>102</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Parallel Corpus for Indigenous Language Translation: Spanish-Mazatec and Spanish-Mixtec
%A Tonja, Atnafu Lambebo
%A Maldonado-sifuentes, Christian
%A Mendoza Castillo, David Alejandro
%A Kolesnikova, Olga
%A Castro-Sánchez, Noé
%A Sidorov, Grigori
%A Gelbukh, Alexander
%Y Mager, Manuel
%Y Ebrahimi, Abteen
%Y Oncevay, Arturo
%Y Rice, Enora
%Y Rijhwani, Shruti
%Y Palmer, Alexis
%Y Kann, Katharina
%S Proceedings of the Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F tonja-etal-2023-parallel
%X In this paper, we present a parallel Spanish- Mazatec and Spanish-Mixtec corpus for machine translation (MT) tasks, where Mazatec and Mixtec are two indigenous Mexican languages. We evaluated the usability of the collected corpus using three different approaches: transformer, transfer learning, and fine-tuning pre-trained multilingual MT models. Fine-tuning the Facebook m2m100-48 model outperformed the other approaches, with BLEU scores of 12.09 and 22.25 for Mazatec-Spanish and Spanish-Mazatec translations, respectively, and 16.75 and 22.15 for Mixtec-Spanish and Spanish-Mixtec translations, respectively. The results indicate that translation performance is influenced by the dataset size (9,799 sentences in Mazatec and 13,235 sentences in Mixtec) and is more effective when indigenous languages are used as target languages. The findings emphasize the importance of creating parallel corpora for indigenous languages and fine-tuning models for low-resource translation tasks. Future research will investigate zero-shot and few-shot learning approaches to further improve translation performance in low-resource settings.
%R 10.18653/v1/2023.americasnlp-1.11
%U https://aclanthology.org/2023.americasnlp-1.11/
%U https://doi.org/10.18653/v1/2023.americasnlp-1.11
%P 94-102
Markdown (Informal)
[Parallel Corpus for Indigenous Language Translation: Spanish-Mazatec and Spanish-Mixtec](https://aclanthology.org/2023.americasnlp-1.11/) (Tonja et al., AmericasNLP 2023)
ACL
- Atnafu Lambebo Tonja, Christian Maldonado-sifuentes, David Alejandro Mendoza Castillo, Olga Kolesnikova, Noé Castro-Sánchez, Grigori Sidorov, and Alexander Gelbukh. 2023. Parallel Corpus for Indigenous Language Translation: Spanish-Mazatec and Spanish-Mixtec. In Proceedings of the Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP), pages 94–102, Toronto, Canada. Association for Computational Linguistics.