What is it about?
Our paper is a survey of multilingual neural machine translation. We focus on summarizing, categorizing and comparing papers that use multilingualism to improve neural machine translation. We have covered over 100 prominent papers from 2016-2020. Our paper will be helpful to beginners as well as experienced machine translation practitioners.
Featured Image
Photo by Leonardo Toshiro Okubo on Unsplash
Why is it important?
Multilingual neural machine translation, ever since its initial success in 2016, has proliferated and there are dozens of papers addressing multilingualism each month. Given the large number of papers published over 2016-2020 we thought that it was important to summarize all works in that time span and create a go-to paper for people to read when they want to get a broad perspective of the field as well as the trends. We have also listed some future directions that multilingual neural machine translation research should/may take and hope that it helps generate new ideas.
Perspectives
When we started writing this article, we planned on a simple summary but over time, thanks to the insightful comments of reviewers our manuscript became longer and more sophisticated. The more we wrote, the more we learned. The process involved a lot of deep thinking and understanding of a variety of ideas but we enjoyed it sufficiently. We hope that our summary will help a number of machine translation practitioners and attract new researchers towards this fascinating (sub-) field of multilingual neural machine translation.
Raj Dabre
Read the Original
This page is a summary of: A Survey of Multilingual Neural Machine Translation, ACM Computing Surveys, October 2020, ACM (Association for Computing Machinery),
DOI: 10.1145/3406095.
You can read the full text:
Resources
Contributors
The following have contributed to this page
