Inducing Language-Agnostic Multilingual Representations

Wei Zhao, Steffen Eger, Johannes Bjerva, Isabelle Augenstein

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

25 Citationer (Scopus)

Abstract

Cross-lingual representations have the potential to make NLP techniques available to the vast majority of languages in the world. However, they currently require large pretraining corpora or access to typologically similar languages. In this work, we address these obstacles by removing language identity signals from multilingual embeddings. We examine three approaches for this: (i) re-aligning the vector spaces of target languages (all together) to a pivot source language; (ii) removing language-specific means and variances, which yields better discriminativeness of embeddings as a by-product; and (iii) increasing input similarity across languages by removing morphological contractions and sentence reordering. We evaluate on XNLI and reference-free MT across 19 typologically diverse languages. Our findings expose the limitations of these approaches-unlike vector normalization, vector space re-alignment and text normalization do not achieve consistent gains across encoders and languages. Due to the approaches' additive effects, their combination decreases the cross-lingual transfer gap by 8.9 points (m-BERT) and 18.2 points (XLM-R) on average across all tasks and languages, however. Our code and models are publicly available.

OriginalsprogEngelsk
Titel*SEM 2021 - 10th Conference on Lexical and Computational Semantics, Proceedings of the Conference
RedaktørerLun-Wei Ku, Vivi Nastase, Ivan Vulic
Antal sider12
ForlagAssociation for Computational Linguistics, ACL Anthology
Publikationsdato2021
Sider229-240
ISBN (Elektronisk)9781954085770
StatusUdgivet - 2021
Begivenhed10th Conference on Lexical and Computational Semantics, *SEM 2021 - Virtual, Bangkok, Thailand
Varighed: 5 aug. 20216 aug. 2021

Konference

Konference10th Conference on Lexical and Computational Semantics, *SEM 2021
Land/OmrådeThailand
ByVirtual, Bangkok
Periode05/08/202106/08/2021
SponsorACL Special Interest Group on the Lexicon, SIGLEX
Navn*SEM 2021 - 10th Conference on Lexical and Computational Semantics, Proceedings of the Conference

Bibliografisk note

Publisher Copyright:
© 2021 Lexical and Computational Semantics

Fingeraftryk

Dyk ned i forskningsemnerne om 'Inducing Language-Agnostic Multilingual Representations'. Sammen danner de et unikt fingeraftryk.

Citationsformater