Overcoming the Rare Word Problem for Low-Resource Language Pairs in Neural Machine Translation

TL;DR

This paper introduces three novel methods to mitigate the rare word problem in neural machine translation for low-resource languages, improving translation quality by up to 1 BLEU point.

Contribution

The paper presents three innovative solutions: enhancing source context, learning morphology of unknown words, and leveraging WordNet synonyms to address rare words in NMT.

Findings

Achieved up to +1.0 BLEU point improvement

Effective in low-resource English-Vietnamese and Japanese-Vietnamese translation

Demonstrated significant reduction in rare word errors

Abstract

Among the six challenges of neural machine translation (NMT) coined by (Koehn and Knowles, 2017), rare-word problem is considered the most severe one, especially in translation of low-resource languages. In this paper, we propose three solutions to address the rare words in neural machine translation systems. First, we enhance source context to predict the target words by connecting directly the source embeddings to the output of the attention component in NMT. Second, we propose an algorithm to learn morphology of unknown words for English in supervised way in order to minimize the adverse effect of rare-word problem. Finally, we exploit synonymous relation from the WordNet to overcome out-of-vocabulary (OOV) problem of NMT. We evaluate our approaches on two low-resource language pairs: English-Vietnamese and Japanese-Vietnamese. In our experiments, we have achieved significant improvements of up to roughly +1.0 BLEU points in both language pairs.