Work

Monolingual Adapters for Zero-Shot Neural Machine Translation

Year: 2020

Type: conference-paper

Abstract: We propose a novel adapter layer formalism for adapting multilingual models.They are more parameter-efficient than existing adapter layers while obtaining as good or better performance.The layers are specific to one language (as opposed to bilingual adapters) allowing to compose them and generalize to unseen language-pairs.In this zero-shot setting, they obtain a median improvement of +2.77BLEU points over a strong 20-language multilingual Transformer baseline trained on TED talks.

Source: Conference on Empirical Methods in Natural Language Processing (EMNLP)

DOI: 10.18653/v1/2020.emnlp-main.361

Authors: Jerin Philip, Alexandre Berard, Matthias Gallé, Laurent Besacier

Institutions: Naver (South Korea), Université Grenoble Alpes

Language: English

Topic: Natural Language Processing Techniques

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W3096966601

FWCI: 4.866

Cites: 22

Cited by: 62

Related to: 10