Robust Machine Translation with Domain Sensitive Pseudo-Sources: Baidu-OSU WMT19 MT Robustness Shared Task System Report

TL;DR

This paper presents a domain-sensitive training approach for robust machine translation of social media text, utilizing pseudo-noisy sources and limited social media data to significantly improve translation quality.

Contribution

The paper introduces a novel domain-sensitive training method that combines real and pseudo-noisy data to enhance social media translation robustness.

Findings

Over 10 BLEU score improvement in En-Fr and Fr-En translation.

Effective use of pseudo-noisy sources generated from monolingual data.

Demonstrated robustness in social media translation tasks.

Abstract

This paper describes the machine translation system developed jointly by Baidu Research and Oregon State University for WMT 2019 Machine Translation Robustness Shared Task. Translation of social media is a very challenging problem, since its style is very different from normal parallel corpora (e.g. News) and also include various types of noises. To make it worse, the amount of social media parallel corpora is extremely limited. In this paper, we use a domain sensitive training method which leverages a large amount of parallel data from popular domains together with a little amount of parallel data from social media. Furthermore, we generate a parallel dataset with pseudo noisy source sentences which are back-translated from monolingual data using a model trained by a similar domain sensitive way. We achieve more than 10 BLEU improvement in both En-Fr and Fr-En translation compared with the baseline methods.