Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction

Yova Radoslavova Kementchedjhieva; Sebastian  Ruder; Ryan  Cotterell; Anders Søgaard

Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction

Yova Radoslavova Kementchedjhieva, Sebastian Ruder, Ryan Cotterell, Anders Søgaard

15 Citationer (Scopus)

Abstract

Most recent approaches to bilingual dictionaryinduction find a linear alignment between theword vector spaces of two languages. Weshow that projecting the two languages ontoa third, latent space, rather than directly ontoeach other, while equivalent in terms of expressivity,makes it easier to learn approximatealignments. Our modified approach also allowsfor supporting languages to be includedin the alignment process, to obtain an even betterperformance in low resource settings.

Originalsprog	Engelsk
Titel	Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018)
Forlag	Association for Computational Linguistics
Publikationsdato	2018
Sider	211–220
Status	Udgivet - 2018
Begivenhed	22nd Conference on Computational Natural Language Learning (CoNLL 2018) - Brussels, Belgien Varighed: 31 okt. 2018 → 1 nov. 2018

Konference

Konference	22nd Conference on Computational Natural Language Learning (CoNLL 2018)
Land/Område	Belgien
By	Brussels
Periode	31/10/2018 → 01/11/2018

Citationsformater

Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction. / Kementchedjhieva, Yova Radoslavova; Ruder, Sebastian ; Cotterell, Ryan et al.

Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018). Association for Computational Linguistics, 2018. s. 211–220.

Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › peer review

Kementchedjhieva, YR, Ruder, S, Cotterell, R & Søgaard, A 2018, Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction. i Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018). Association for Computational Linguistics, s. 211–220, 22nd Conference on Computational Natural Language Learning (CoNLL 2018), Brussels, Belgien, 31/10/2018.

@inproceedings{3a8ec48814ff4a6cbc086cdfa14cd47e,

title = "Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction",

abstract = "Most recent approaches to bilingual dictionaryinduction find a linear alignment between theword vector spaces of two languages. Weshow that projecting the two languages ontoa third, latent space, rather than directly ontoeach other, while equivalent in terms of expressivity,makes it easier to learn approximatealignments. Our modified approach also allowsfor supporting languages to be includedin the alignment process, to obtain an even betterperformance in low resource settings.",

author = "Kementchedjhieva, {Yova Radoslavova} and Sebastian Ruder and Ryan Cotterell and Anders S{\o}gaard",

year = "2018",

language = "English",

pages = "211–220",

booktitle = "Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018)",

publisher = "Association for Computational Linguistics",

note = "22nd Conference on Computational Natural Language Learning (CoNLL 2018) ; Conference date: 31-10-2018 Through 01-11-2018",

}

TY - GEN

T1 - Generalizing Procrustes Analysis for Better Bilingual Dictionary Induction

AU - Kementchedjhieva, Yova Radoslavova

AU - Ruder, Sebastian

AU - Cotterell, Ryan

AU - Søgaard, Anders

PY - 2018

Y1 - 2018

N2 - Most recent approaches to bilingual dictionaryinduction find a linear alignment between theword vector spaces of two languages. Weshow that projecting the two languages ontoa third, latent space, rather than directly ontoeach other, while equivalent in terms of expressivity,makes it easier to learn approximatealignments. Our modified approach also allowsfor supporting languages to be includedin the alignment process, to obtain an even betterperformance in low resource settings.

AB - Most recent approaches to bilingual dictionaryinduction find a linear alignment between theword vector spaces of two languages. Weshow that projecting the two languages ontoa third, latent space, rather than directly ontoeach other, while equivalent in terms of expressivity,makes it easier to learn approximatealignments. Our modified approach also allowsfor supporting languages to be includedin the alignment process, to obtain an even betterperformance in low resource settings.

M3 - Article in proceedings

SP - 211

EP - 220

BT - Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018)

PB - Association for Computational Linguistics

T2 - 22nd Conference on Computational Natural Language Learning (CoNLL 2018)

Y2 - 31 October 2018 through 1 November 2018

ER -