Multilingual Machine Translation: Deep Analysis of Language-Specific Encoder-Decoders

Carlos Escolano, Marta Ruiz Costa-jussà, José A. R. Fonollosa

JAIR 2022 pp. 1535-1552

doi:10.1613/JAIR.1.12699 /jair/2022/escolano2022jair-multilingual/

Abstract

State-of-the-art multilingual machine translation relies on a shared encoder-decoder. In this paper, we propose an alternative approach based on language-specific encoder-decoders, which can be easily extended to new languages by learning their corresponding modules. To establish a common interlingua representation, we simultaneously train N initial languages. Our experiments show that the proposed approach improves over the shared encoder-decoder for the initial languages and when adding new languages, without the need to retrain the remaining modules. All in all, our work closes the gap between shared and language-specific encoder-decoders, advancing toward modular multilingual machine translation systems that can be flexibly extended in lifelong learning settings.

PDF JAIR Semantic Scholar

Cite

Text

Escolano et al. "Multilingual Machine Translation: Deep Analysis of Language-Specific Encoder-Decoders." Journal of Artificial Intelligence Research, 2022. doi:10.1613/JAIR.1.12699

Markdown

[Escolano et al. "Multilingual Machine Translation: Deep Analysis of Language-Specific Encoder-Decoders." Journal of Artificial Intelligence Research, 2022.](https://mlanthology.org/jair/2022/escolano2022jair-multilingual/) doi:10.1613/JAIR.1.12699

BibTeX

@article{escolano2022jair-multilingual,
  title     = {{Multilingual Machine Translation: Deep Analysis of Language-Specific Encoder-Decoders}},
  author    = {Escolano, Carlos and Costa-jussà, Marta Ruiz and Fonollosa, José A. R.},
  journal   = {Journal of Artificial Intelligence Research},
  year      = {2022},
  pages     = {1535-1552},
  doi       = {10.1613/JAIR.1.12699},
  volume    = {73},
  url       = {https://mlanthology.org/jair/2022/escolano2022jair-multilingual/}
}