Multi-Domain Adaptation in Neural Machine Translation Through Multidimensional Tagging

Emmanouil Stergiadis, Satendra Kumar, Fedor Kovalev, Pavel Levin


Abstract
Production NMT systems typically need to serve niche domains that are not covered by adequately large and readily available parallel corpora. As a result, practitioners often fine-tune general purpose models to each of the domains their organisation caters to. The number of domains however can often become large, which in combination with the number of languages that need serving can lead to an unscalable fleet of models to be developed and maintained. We propose Multi Dimensional Tagging, a method for fine-tuning a single NMT model on several domains simultaneously, thus drastically reducing development and maintenance costs. We run experiments where a single MDT model compares favourably to a set of SOTA specialist models, even when evaluated on the domain those baselines have been fine-tuned on. Besides BLEU, we report human evaluation results. MDT models are now live at Booking.com, powering an MT engine that serves millions of translations a day in over 40 different languages.
Anthology ID:
2021.mtsummit-up.27
Volume:
Proceedings of Machine Translation Summit XVIII: Users and Providers Track
Month:
August
Year:
2021
Address:
Virtual
Editors:
Janice Campbell, Ben Huyck, Stephen Larocca, Jay Marciano, Konstantin Savenkov, Alex Yanishevsky
Venue:
MTSummit
SIG:
Publisher:
Association for Machine Translation in the Americas
Note:
Pages:
396–420
Language:
URL:
https://aclanthology.org/2021.mtsummit-up.27
DOI:
Bibkey:
Cite (ACL):
Emmanouil Stergiadis, Satendra Kumar, Fedor Kovalev, and Pavel Levin. 2021. Multi-Domain Adaptation in Neural Machine Translation Through Multidimensional Tagging. In Proceedings of Machine Translation Summit XVIII: Users and Providers Track, pages 396–420, Virtual. Association for Machine Translation in the Americas.
Cite (Informal):
Multi-Domain Adaptation in Neural Machine Translation Through Multidimensional Tagging (Stergiadis et al., MTSummit 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.mtsummit-up.27.pdf
Presentation:
 2021.mtsummit-up.27.Presentation.pdf
Code
 emnlpanon/mt-paper