Synthetic Source Language Augmentation for Colloquial Neural Machine Translation

TL;DR

This paper introduces a synthetic style augmentation method for NMT that enhances translation of colloquial Indonesian by creating a new test set and improving model performance on informal language.

Contribution

The work develops a novel colloquial Indonesian-English test set and demonstrates that synthetic style augmentation of formal source data improves NMT performance on colloquial language.

Findings

Improved BLEU scores on colloquial Indonesian-English translation

Created a new colloquial Indonesian-English test set from YouTube and Twitter

Synthetic style augmentation benefits NMT in handling informal language

Abstract

Neural machine translation (NMT) is typically domain-dependent and style-dependent, and it requires lots of training data. State-of-the-art NMT models often fall short in handling colloquial variations of its source language and the lack of parallel data in this regard is a challenging hurdle in systematically improving the existing models. In this work, we develop a novel colloquial Indonesian-English test-set collected from YouTube transcript and Twitter. We perform synthetic style augmentation to the source of formal Indonesian language and show that it improves the baseline Id-En models (in BLEU) over the new test data.