Linguistic Input Features Improve Neural Machine Translation

TL;DR

Incorporating linguistic features such as morphology, POS tags, and syntactic dependencies into neural machine translation models enhances translation quality across multiple language pairs, despite the models' strong baseline performance.

Contribution

We extend the encoder's embedding layer to support arbitrary linguistic features, demonstrating their positive impact on neural MT performance.

Findings

Linguistic features improve perplexity, BLEU, and CHRF3 scores.

Adding linguistic features yields consistent gains across language pairs.

Open-source implementation and sample configurations are provided.

Abstract

Neural machine translation has recently achieved impressive results, while using little in the way of external linguistic information. In this paper we show that the strong learning capability of neural MT models does not make linguistic features redundant; they can be easily incorporated to provide further improvements in performance. We generalize the embedding layer of the encoder in the attentional encoder--decoder architecture to support the inclusion of arbitrary features, in addition to the baseline word feature. We add morphological features, part-of-speech tags, and syntactic dependency labels as input features to English<->German, and English->Romanian neural machine translation systems. In experiments on WMT16 training and test sets, we find that linguistic input features improve model quality according to three metrics: perplexity, BLEU and CHRF3. An open-source implementation of our neural MT system is available, as are sample files and configurations.