Multilingual training of crosslingual word embeddings

Long Duong; Hiroshi Kanayama; Tengfei Ma; Steven Bird; Trevor Cohn

doi:10.18653/v1/e17-1084

Multilingual training of crosslingual word embeddings

Long Duong
, Hiroshi Kanayama
, Tengfei Ma
, Steven Bird
, Trevor Cohn

Biomedical Informatics

University of Melbourne
IBM
University of California at Berkeley

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

36 Scopus citations

Abstract

Crosslingual word embeddings represent lexical items from different languages using the same vector space, enabling crosslingual transfer. Most prior work constructs embeddings for a pair of languages, with English on one side. We investigate methods for building high quality crosslingual word embeddings for many languages in a unified vector space. In this way, we can exploit and combine information from many languages. We report competitive performance on bilingual lexicon induction, monolingual similarity and crosslingual document classification tasks.

Original language	English
Title of host publication	Long Papers - Continued
Publisher	Association for Computational Linguistics (ACL)
Pages	894-904
Number of pages	11
ISBN (Electronic)	9781510838604
DOIs	https://doi.org/10.18653/v1/e17-1084
State	Published - Jul 1 2017
Event	15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Valencia, Spain Duration: Apr 3 2017 → Apr 7 2017

Publication series

Name	15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference
Volume	1

Conference

Conference	15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017
Country/Territory	Spain
City	Valencia
Period	04/3/17 → 04/7/17

Access to Document

10.18653/v1/e17-1084

Cite this

Duong, L., Kanayama, H., Ma, T., Bird, S., & Cohn, T. (2017). Multilingual training of crosslingual word embeddings. In Long Papers - Continued (pp. 894-904). (15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference; Vol. 1). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/e17-1084

@inproceedings{45db174423854b6a9083bcff748d1bed,

title = "Multilingual training of crosslingual word embeddings",

abstract = "Crosslingual word embeddings represent lexical items from different languages using the same vector space, enabling crosslingual transfer. Most prior work constructs embeddings for a pair of languages, with English on one side. We investigate methods for building high quality crosslingual word embeddings for many languages in a unified vector space. In this way, we can exploit and combine information from many languages. We report competitive performance on bilingual lexicon induction, monolingual similarity and crosslingual document classification tasks.",

author = "Long Duong and Hiroshi Kanayama and Tengfei Ma and Steven Bird and Trevor Cohn",

note = "Publisher Copyright: {\textcopyright} 2017 Association for Computational Linguistics.; 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 ; Conference date: 03-04-2017 Through 07-04-2017",

year = "2017",

month = jul,

day = "1",

doi = "10.18653/v1/e17-1084",

language = "English",

series = "15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference",

publisher = "Association for Computational Linguistics (ACL)",

pages = "894--904",

booktitle = "Long Papers - Continued",

}

Duong, L, Kanayama, H, Ma, T, Bird, S & Cohn, T 2017, Multilingual training of crosslingual word embeddings. in Long Papers - Continued. 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference, vol. 1, Association for Computational Linguistics (ACL), pp. 894-904, 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017, Valencia, Spain, 04/3/17. https://doi.org/10.18653/v1/e17-1084

Multilingual training of crosslingual word embeddings. / Duong, Long; Kanayama, Hiroshi; Ma, Tengfei et al.
Long Papers - Continued. Association for Computational Linguistics (ACL), 2017. p. 894-904 (15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference; Vol. 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Multilingual training of crosslingual word embeddings

AU - Duong, Long

AU - Kanayama, Hiroshi

AU - Ma, Tengfei

AU - Bird, Steven

AU - Cohn, Trevor

PY - 2017/7/1

Y1 - 2017/7/1

N2 - Crosslingual word embeddings represent lexical items from different languages using the same vector space, enabling crosslingual transfer. Most prior work constructs embeddings for a pair of languages, with English on one side. We investigate methods for building high quality crosslingual word embeddings for many languages in a unified vector space. In this way, we can exploit and combine information from many languages. We report competitive performance on bilingual lexicon induction, monolingual similarity and crosslingual document classification tasks.

AB - Crosslingual word embeddings represent lexical items from different languages using the same vector space, enabling crosslingual transfer. Most prior work constructs embeddings for a pair of languages, with English on one side. We investigate methods for building high quality crosslingual word embeddings for many languages in a unified vector space. In this way, we can exploit and combine information from many languages. We report competitive performance on bilingual lexicon induction, monolingual similarity and crosslingual document classification tasks.

UR - https://www.scopus.com/pages/publications/85021686106

U2 - 10.18653/v1/e17-1084

DO - 10.18653/v1/e17-1084

M3 - Conference contribution

AN - SCOPUS:85021686106

T3 - 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017 - Proceedings of Conference

SP - 894

EP - 904

BT - Long Papers - Continued

PB - Association for Computational Linguistics (ACL)

T2 - 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017

Y2 - 3 April 2017 through 7 April 2017

ER -

Multilingual training of crosslingual word embeddings

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this