Neural machine translation for low-resource languages

TL;DR

This paper explores neural machine translation for low-resource languages by introducing local dependencies and word alignments, showing it can produce acceptable translations with limited data, though SMT still performs better in such scenarios.

Contribution

The paper presents a novel NMT model tailored for low-resource languages and compares its performance with traditional SMT in low-data conditions.

Findings

NMT can produce acceptable translations with 70,000 tokens of data.

SMT outperforms NMT in very low-resource settings.

The proposed NMT model incorporates local dependencies and word alignments.

Abstract

Neural machine translation (NMT) approaches have improved the state of the art in many machine translation settings over the last couple of years, but they require large amounts of training data to produce sensible output. We demonstrate that NMT can be used for low-resource languages as well, by introducing more local dependencies and using word alignments to learn sentence reordering during translation. In addition to our novel model, we also present an empirical evaluation of low-resource phrase-based statistical machine translation (SMT) and NMT to investigate the lower limits of the respective technologies. We find that while SMT remains the best option for low-resource settings, our method can produce acceptable translations with only 70000 tokens of training data, a level where the baseline NMT system fails completely.