@inproceedings{li-etal-2023-multilingual,
title = "Multilingual Bidirectional Unsupervised Translation through Multilingual Finetuning and Back-Translation",
author = "Li, Bryan and
Rasooli, Mohammad Sadegh and
Patel, Ajay and
Callison-burch, Chris",
editor = "Ojha, Atul Kr. and
Liu, Chao-hong and
Vylomova, Ekaterina and
Pirinen, Flammie and
Abbott, Jade and
Washington, Jonathan and
Oco, Nathaniel and
Malykh, Valentin and
Logacheva, Varvara and
Zhao, Xiaobing",
booktitle = "Proceedings of the Sixth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2023)",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.loresmt-1.2",
doi = "10.18653/v1/2023.loresmt-1.2",
pages = "16--31",
abstract = "We propose a two-stage approach for training a single NMT model to translate unseen languages both to and from English. For the first stage, we initialize an encoder-decoder model to pretrained XLM-R and RoBERTa weights, then perform multilingual fine-tuning on parallel data in 40 languages to English. We find this model can generalize to zero-shot translations on unseen languages. For the second stage, we leverage this generalization ability to generate synthetic parallel data from monolingual datasets, then bidirectionally train with successive rounds of back-translation. Our approach, which we EcXTra (uE/unglish-uc/uentric Crosslingual (uX/u) uTra/unsfer), is conceptually simple, only using a standard cross-entropy objective throughout. It is also data-driven, sequentially leveraging auxiliary parallel data and monolingual data. We evaluate unsupervised NMT results for 7 low-resource languages, and find that each round of back-translation training further refines bidirectional performance. Our final single EcXTra-trained model achieves competitive translation performance in all translation directions, notably establishing a new state-of-the-art for English-to-Kazakh (22.9 10.4 BLEU). Our code is available at [this URL](\url{https://github.com/manestay/EcXTra}).",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2023-multilingual">
<titleInfo>
<title>Multilingual Bidirectional Unsupervised Translation through Multilingual Finetuning and Back-Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bryan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Sadegh</namePart>
<namePart type="family">Rasooli</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ajay</namePart>
<namePart type="family">Patel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chris</namePart>
<namePart type="family">Callison-burch</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Sixth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2023)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Atul</namePart>
<namePart type="given">Kr.</namePart>
<namePart type="family">Ojha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chao-hong</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Vylomova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Flammie</namePart>
<namePart type="family">Pirinen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jade</namePart>
<namePart type="family">Abbott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jonathan</namePart>
<namePart type="family">Washington</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nathaniel</namePart>
<namePart type="family">Oco</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Valentin</namePart>
<namePart type="family">Malykh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Varvara</namePart>
<namePart type="family">Logacheva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaobing</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dubrovnik, Croatia</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We propose a two-stage approach for training a single NMT model to translate unseen languages both to and from English. For the first stage, we initialize an encoder-decoder model to pretrained XLM-R and RoBERTa weights, then perform multilingual fine-tuning on parallel data in 40 languages to English. We find this model can generalize to zero-shot translations on unseen languages. For the second stage, we leverage this generalization ability to generate synthetic parallel data from monolingual datasets, then bidirectionally train with successive rounds of back-translation. Our approach, which we EcXTra (uE/unglish-uc/uentric Crosslingual (uX/u) uTra/unsfer), is conceptually simple, only using a standard cross-entropy objective throughout. It is also data-driven, sequentially leveraging auxiliary parallel data and monolingual data. We evaluate unsupervised NMT results for 7 low-resource languages, and find that each round of back-translation training further refines bidirectional performance. Our final single EcXTra-trained model achieves competitive translation performance in all translation directions, notably establishing a new state-of-the-art for English-to-Kazakh (22.9 10.4 BLEU). Our code is available at [this URL](https://github.com/manestay/EcXTra).</abstract>
<identifier type="citekey">li-etal-2023-multilingual</identifier>
<identifier type="doi">10.18653/v1/2023.loresmt-1.2</identifier>
<location>
<url>https://aclanthology.org/2023.loresmt-1.2</url>
</location>
<part>
<date>2023-05</date>
<extent unit="page">
<start>16</start>
<end>31</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Multilingual Bidirectional Unsupervised Translation through Multilingual Finetuning and Back-Translation
%A Li, Bryan
%A Rasooli, Mohammad Sadegh
%A Patel, Ajay
%A Callison-burch, Chris
%Y Ojha, Atul Kr.
%Y Liu, Chao-hong
%Y Vylomova, Ekaterina
%Y Pirinen, Flammie
%Y Abbott, Jade
%Y Washington, Jonathan
%Y Oco, Nathaniel
%Y Malykh, Valentin
%Y Logacheva, Varvara
%Y Zhao, Xiaobing
%S Proceedings of the Sixth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2023)
%D 2023
%8 May
%I Association for Computational Linguistics
%C Dubrovnik, Croatia
%F li-etal-2023-multilingual
%X We propose a two-stage approach for training a single NMT model to translate unseen languages both to and from English. For the first stage, we initialize an encoder-decoder model to pretrained XLM-R and RoBERTa weights, then perform multilingual fine-tuning on parallel data in 40 languages to English. We find this model can generalize to zero-shot translations on unseen languages. For the second stage, we leverage this generalization ability to generate synthetic parallel data from monolingual datasets, then bidirectionally train with successive rounds of back-translation. Our approach, which we EcXTra (uE/unglish-uc/uentric Crosslingual (uX/u) uTra/unsfer), is conceptually simple, only using a standard cross-entropy objective throughout. It is also data-driven, sequentially leveraging auxiliary parallel data and monolingual data. We evaluate unsupervised NMT results for 7 low-resource languages, and find that each round of back-translation training further refines bidirectional performance. Our final single EcXTra-trained model achieves competitive translation performance in all translation directions, notably establishing a new state-of-the-art for English-to-Kazakh (22.9 10.4 BLEU). Our code is available at [this URL](https://github.com/manestay/EcXTra).
%R 10.18653/v1/2023.loresmt-1.2
%U https://aclanthology.org/2023.loresmt-1.2
%U https://doi.org/10.18653/v1/2023.loresmt-1.2
%P 16-31
Markdown (Informal)
[Multilingual Bidirectional Unsupervised Translation through Multilingual Finetuning and Back-Translation](https://aclanthology.org/2023.loresmt-1.2) (Li et al., LoResMT 2023)
ACL