True Bilingual NMT

Mohamed Anwar, Lekan Raheem, Maab Elrasheed, Melvin Johnson, Julia Kreutzer

ICLRW 2022

/iclrw/2022/anwar2022iclrw-true/

Abstract

Bilingual machine translation permits training a single model that translates monolingual sentences from one language to another. However, a model is not truly bilingual unless it can translate back and forth in both language directions it was trained on, along with translating code-switched sentences to either language. We propose a true bilingual model trained on WMT14 English-French (En-Fr) dataset. For better use of parallel data, we generated synthetic code-switched (CSW) data along with an alignment loss on the encoder to align representations across languages. Our model strongly outperforms bilingual baselines on CSW translation while maintaining quality for non-code switched data.

PDF ICLRW OpenReview Semantic Scholar

Cite

Text

Anwar et al. "True Bilingual NMT." ICLR 2022 Workshops: AfricaNLP, 2022.

Markdown

[Anwar et al. "True Bilingual NMT." ICLR 2022 Workshops: AfricaNLP, 2022.](https://mlanthology.org/iclrw/2022/anwar2022iclrw-true/)

BibTeX

@inproceedings{anwar2022iclrw-true,
  title     = {{True Bilingual NMT}},
  author    = {Anwar, Mohamed and Raheem, Lekan and Elrasheed, Maab and Johnson, Melvin and Kreutzer, Julia},
  booktitle = {ICLR 2022 Workshops: AfricaNLP},
  year      = {2022},
  url       = {https://mlanthology.org/iclrw/2022/anwar2022iclrw-true/}
}