@inproceedings{downey-etal-2022-multilingual,
title = "Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages",
author = "Downey, C. and
Drizin, Shannon and
Haroutunian, Levon and
Thukral, Shivin",
editor = "Muresan, Smaranda and
Nakov, Preslav and
Villavicencio, Aline",
booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = may,
year = "2022",
address = "Dublin, Ireland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.acl-long.366/",
doi = "10.18653/v1/2022.acl-long.366",
pages = "5331--5346",
abstract = "We show that unsupervised sequence-segmentation performance can be transferred to extremely low-resource languages by pre-training a Masked Segmental Language Model (Downey et al., 2021) multilingually. Further, we show that this transfer can be achieved by training over a collection of low-resource languages that are typologically similar (but phylogenetically unrelated) to the target language. In our experiments, we transfer from a collection of 10 Indigenous American languages (AmericasNLP, Mager et al., 2021) to K`iche', a Mayan language. We compare our multilingual model to a monolingual (from-scratch) baseline, as well as a model pre-trained on Quechua only. We show that the multilingual pre-trained approach yields consistent segmentation quality across target dataset sizes, exceeding the monolingual baseline in 6/10 experimental settings. Our model yields especially strong results at small target sizes, including a zero-shot performance of 20.6 F1. These results have promising implications for low-resource NLP pipelines involving human-like linguistic units, such as the sparse transcription framework proposed by Bird (2020)."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="downey-etal-2022-multilingual">
<titleInfo>
<title>Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages</title>
</titleInfo>
<name type="personal">
<namePart type="given">C</namePart>
<namePart type="family">Downey</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shannon</namePart>
<namePart type="family">Drizin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Levon</namePart>
<namePart type="family">Haroutunian</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shivin</namePart>
<namePart type="family">Thukral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Smaranda</namePart>
<namePart type="family">Muresan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Preslav</namePart>
<namePart type="family">Nakov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aline</namePart>
<namePart type="family">Villavicencio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dublin, Ireland</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We show that unsupervised sequence-segmentation performance can be transferred to extremely low-resource languages by pre-training a Masked Segmental Language Model (Downey et al., 2021) multilingually. Further, we show that this transfer can be achieved by training over a collection of low-resource languages that are typologically similar (but phylogenetically unrelated) to the target language. In our experiments, we transfer from a collection of 10 Indigenous American languages (AmericasNLP, Mager et al., 2021) to K‘iche’, a Mayan language. We compare our multilingual model to a monolingual (from-scratch) baseline, as well as a model pre-trained on Quechua only. We show that the multilingual pre-trained approach yields consistent segmentation quality across target dataset sizes, exceeding the monolingual baseline in 6/10 experimental settings. Our model yields especially strong results at small target sizes, including a zero-shot performance of 20.6 F1. These results have promising implications for low-resource NLP pipelines involving human-like linguistic units, such as the sparse transcription framework proposed by Bird (2020).</abstract>
<identifier type="citekey">downey-etal-2022-multilingual</identifier>
<identifier type="doi">10.18653/v1/2022.acl-long.366</identifier>
<location>
<url>https://aclanthology.org/2022.acl-long.366/</url>
</location>
<part>
<date>2022-05</date>
<extent unit="page">
<start>5331</start>
<end>5346</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages
%A Downey, C.
%A Drizin, Shannon
%A Haroutunian, Levon
%A Thukral, Shivin
%Y Muresan, Smaranda
%Y Nakov, Preslav
%Y Villavicencio, Aline
%S Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2022
%8 May
%I Association for Computational Linguistics
%C Dublin, Ireland
%F downey-etal-2022-multilingual
%X We show that unsupervised sequence-segmentation performance can be transferred to extremely low-resource languages by pre-training a Masked Segmental Language Model (Downey et al., 2021) multilingually. Further, we show that this transfer can be achieved by training over a collection of low-resource languages that are typologically similar (but phylogenetically unrelated) to the target language. In our experiments, we transfer from a collection of 10 Indigenous American languages (AmericasNLP, Mager et al., 2021) to K‘iche’, a Mayan language. We compare our multilingual model to a monolingual (from-scratch) baseline, as well as a model pre-trained on Quechua only. We show that the multilingual pre-trained approach yields consistent segmentation quality across target dataset sizes, exceeding the monolingual baseline in 6/10 experimental settings. Our model yields especially strong results at small target sizes, including a zero-shot performance of 20.6 F1. These results have promising implications for low-resource NLP pipelines involving human-like linguistic units, such as the sparse transcription framework proposed by Bird (2020).
%R 10.18653/v1/2022.acl-long.366
%U https://aclanthology.org/2022.acl-long.366/
%U https://doi.org/10.18653/v1/2022.acl-long.366
%P 5331-5346
Markdown (Informal)
[Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages](https://aclanthology.org/2022.acl-long.366/) (Downey et al., ACL 2022)
ACL