Findings of the LoResMT 2020 Shared Task on Zero-Shot for Low-Resource languages

Atul Kumar Ojha

doi:http://hdl.handle.net/10379/16376

Findings of the LoResMT 2020 Shared Task on Zero-Shot for Low-Resource languages

Research output: Chapter in Book or Conference Publication/Proceeding › Conference Publication › peer-review

Abstract

This paper presents the findings of the LoResMT 2020 Shared Task on zero-shot translation for low resource languages. This task was organised as part of the 3rd Workshop on Technologies for MT of Low Resource Languages (LoResMT) at AACL-IJCNLP 2020. The focus was on the zero-shot approach as a notable development in Neural Machine Translation to build MT systems for language pairs where parallel corpora are small or even non-existent. The shared task experience suggests that back-translation and domain adaptation methods result in better accuracy for small-size datasets. We further noted that, although translation between similar languages is no cakewalk, linguistically distinct languages require more data to give better results.

Original language	English (Ireland)
Title of host publication	Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages
Place of Publication	Suzhou, China
DOIs	https://doi.org/http://hdl.handle.net/10379/16376
Publication status	Published - 1 Dec 2020

Authors (Note for portal: view the doc link for the full list of authors)

Authors
Ojha, Atul Kr.; Malykh, Valentin; Karakanta, Alina and Liu, Chao-Hong

Access to Document

http://hdl.handle.net/10379/16376

Cite this

@inproceedings{4389617d37ea4fd9bfd77109e7035dc3,

title = "Findings of the LoResMT 2020 Shared Task on Zero-Shot for Low-Resource languages",

abstract = "This paper presents the findings of the LoResMT 2020 Shared Task on zero-shot translation for low resource languages. This task was organised as part of the 3rd Workshop on Technologies for MT of Low Resource Languages (LoResMT) at AACL-IJCNLP 2020. The focus was on the zero-shot approach as a notable development in Neural Machine Translation to build MT systems for language pairs where parallel corpora are small or even non-existent. The shared task experience suggests that back-translation and domain adaptation methods result in better accuracy for small-size datasets. We further noted that, although translation between similar languages is no cakewalk, linguistically distinct languages require more data to give better results.",

author = "Ojha, \{Atul Kumar\}",

year = "2020",

month = dec,

day = "1",

doi = "http://hdl.handle.net/10379/16376",

language = "English (Ireland)",

booktitle = "Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages",

}

TY - GEN

T1 - Findings of the LoResMT 2020 Shared Task on Zero-Shot for Low-Resource languages

AU - Ojha, Atul Kumar

PY - 2020/12/1

Y1 - 2020/12/1

N2 - This paper presents the findings of the LoResMT 2020 Shared Task on zero-shot translation for low resource languages. This task was organised as part of the 3rd Workshop on Technologies for MT of Low Resource Languages (LoResMT) at AACL-IJCNLP 2020. The focus was on the zero-shot approach as a notable development in Neural Machine Translation to build MT systems for language pairs where parallel corpora are small or even non-existent. The shared task experience suggests that back-translation and domain adaptation methods result in better accuracy for small-size datasets. We further noted that, although translation between similar languages is no cakewalk, linguistically distinct languages require more data to give better results.

AB - This paper presents the findings of the LoResMT 2020 Shared Task on zero-shot translation for low resource languages. This task was organised as part of the 3rd Workshop on Technologies for MT of Low Resource Languages (LoResMT) at AACL-IJCNLP 2020. The focus was on the zero-shot approach as a notable development in Neural Machine Translation to build MT systems for language pairs where parallel corpora are small or even non-existent. The shared task experience suggests that back-translation and domain adaptation methods result in better accuracy for small-size datasets. We further noted that, although translation between similar languages is no cakewalk, linguistically distinct languages require more data to give better results.

U2 - http://hdl.handle.net/10379/16376

DO - http://hdl.handle.net/10379/16376

M3 - Conference Publication

BT - Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages

CY - Suzhou, China

ER -

Findings of the LoResMT 2020 Shared Task on Zero-Shot for Low-Resource languages

Abstract

Authors (Note for portal: view the doc link for the full list of authors)

Access to Document

Fingerprint

Cite this