@inproceedings{chai-etal-2022-cross,
title = "Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure",
author = "Chai, Yuan and
Liang, Yaobo and
Duan, Nan",
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.322",
doi = "10.18653/v1/2022.acl-long.322",
pages = "4702--4712",
abstract = "Multilingual pre-trained language models, such as mBERT and XLM-R, have shown impressive cross-lingual ability. Surprisingly, both of them use multilingual masked language model (MLM) without any cross-lingual supervision or aligned data. Despite the encouraging results, we still lack a clear understanding of why cross-lingual ability could emerge from multilingual MLM. In our work, we argue that cross-language ability comes from the commonality between languages. Specifically, we study three language properties: constituent order, composition and word co-occurrence. First, we create an artificial language by modifying property in source language. Then we study the contribution of modified property through the change of cross-language transfer results on target language. We conduct experiments on six languages and two cross-lingual NLP tasks (textual entailment, sentence retrieval). Our main conclusion is that the contribution of constituent order and word co-occurrence is limited, while the composition is more crucial to the success of cross-linguistic transfer.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chai-etal-2022-cross">
<titleInfo>
<title>Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuan</namePart>
<namePart type="family">Chai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yaobo</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nan</namePart>
<namePart type="family">Duan</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>Multilingual pre-trained language models, such as mBERT and XLM-R, have shown impressive cross-lingual ability. Surprisingly, both of them use multilingual masked language model (MLM) without any cross-lingual supervision or aligned data. Despite the encouraging results, we still lack a clear understanding of why cross-lingual ability could emerge from multilingual MLM. In our work, we argue that cross-language ability comes from the commonality between languages. Specifically, we study three language properties: constituent order, composition and word co-occurrence. First, we create an artificial language by modifying property in source language. Then we study the contribution of modified property through the change of cross-language transfer results on target language. We conduct experiments on six languages and two cross-lingual NLP tasks (textual entailment, sentence retrieval). Our main conclusion is that the contribution of constituent order and word co-occurrence is limited, while the composition is more crucial to the success of cross-linguistic transfer.</abstract>
<identifier type="citekey">chai-etal-2022-cross</identifier>
<identifier type="doi">10.18653/v1/2022.acl-long.322</identifier>
<location>
<url>https://aclanthology.org/2022.acl-long.322</url>
</location>
<part>
<date>2022-05</date>
<extent unit="page">
<start>4702</start>
<end>4712</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure
%A Chai, Yuan
%A Liang, Yaobo
%A Duan, Nan
%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 chai-etal-2022-cross
%X Multilingual pre-trained language models, such as mBERT and XLM-R, have shown impressive cross-lingual ability. Surprisingly, both of them use multilingual masked language model (MLM) without any cross-lingual supervision or aligned data. Despite the encouraging results, we still lack a clear understanding of why cross-lingual ability could emerge from multilingual MLM. In our work, we argue that cross-language ability comes from the commonality between languages. Specifically, we study three language properties: constituent order, composition and word co-occurrence. First, we create an artificial language by modifying property in source language. Then we study the contribution of modified property through the change of cross-language transfer results on target language. We conduct experiments on six languages and two cross-lingual NLP tasks (textual entailment, sentence retrieval). Our main conclusion is that the contribution of constituent order and word co-occurrence is limited, while the composition is more crucial to the success of cross-linguistic transfer.
%R 10.18653/v1/2022.acl-long.322
%U https://aclanthology.org/2022.acl-long.322
%U https://doi.org/10.18653/v1/2022.acl-long.322
%P 4702-4712
Markdown (Informal)
[Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure](https://aclanthology.org/2022.acl-long.322) (Chai et al., ACL 2022)
ACL