Improved Neural Machine Translation with a Syntax-Aware Encoder and Decoder

TL;DR

This paper enhances neural machine translation by integrating source-side syntactic trees into the encoder-decoder framework, leading to improved translation quality for Chinese-English tasks.

Contribution

It introduces a bidirectional tree encoder and a tree-coverage model that incorporate syntactic information into NMT, which were not previously used together.

Findings

Outperforms standard sequential NMT models.

Achieves better translation accuracy on Chinese-English translation.

Demonstrates the effectiveness of syntax-aware models.

Abstract

Most neural machine translation (NMT) models are based on the sequential encoder-decoder framework, which makes no use of syntactic information. In this paper, we improve this model by explicitly incorporating source-side syntactic trees. More specifically, we propose (1) a bidirectional tree encoder which learns both sequential and tree structured representations; (2) a tree-coverage model that lets the attention depend on the source-side syntax. Experiments on Chinese-English translation demonstrate that our proposed models outperform the sequential attentional model as well as a stronger baseline with a bottom-up tree encoder and word coverage.