# Towards String-to-Tree Neural Machine Translation

## TL;DR

This paper introduces a syntax-aware neural machine translation approach that translates into linearized constituency trees, improving translation quality and reordering capabilities over syntax-agnostic models, as demonstrated on the WMT16 German-English task.

## Contribution

It proposes a simple method to incorporate syntactic information into NMT by translating into linearized constituency trees, enhancing translation quality and reordering.

## Key findings

- Improved BLEU scores over baseline
- More reordering in translations
- Positive human evaluation results

## Abstract

We present a simple method to incorporate syntactic information about the target language in a neural machine translation system by translating into linearized, lexicalized constituency trees. An experiment on the WMT16 German-English news translation task resulted in an improved BLEU score when compared to a syntax-agnostic NMT baseline trained on the same dataset. An analysis of the translations from the syntax-aware system shows that it performs more reordering during translation in comparison to the baseline. A small-scale human evaluation also showed an advantage to the syntax-aware system.