Modeling Target-Side Inflection in Neural Machine Translation

TL;DR

This paper proposes a method for neural machine translation that improves handling of morphologically rich languages by predicting lemmas and POS tags, leading to better translation quality.

Contribution

It introduces a lemma and POS tagging approach for NMT that enhances generalization over complex morphological systems.

Findings

Improved translation quality for Czech and German.

The approach outperforms standard BPE-based methods.

Explicit morphological information alone does not explain the improvements.

Abstract

NMT systems have problems with large vocabulary sizes. Byte-pair encoding (BPE) is a popular approach to solving this problem, but while BPE allows the system to generate any target-side word, it does not enable effective generalization over the rich vocabulary in morphologically rich languages with strong inflectional phenomena. We introduce a simple approach to overcome this problem by training a system to produce the lemma of a word and its morphologically rich POS tag, which is then followed by a deterministic generation step. We apply this strategy for English-Czech and English-German translation scenarios, obtaining improvements in both settings. We furthermore show that the improvement is not due to only adding explicit morphological information.