Towards Neural Machine Translation with Partially Aligned Corpora

TL;DR

This paper introduces a novel NMT training approach that leverages partially aligned sentence pairs derived from monolingual corpora and phrase pairs, enabling translation in low-resource scenarios.

Contribution

It proposes a new training method for NMT using partially aligned data, adapting generation strategies and objective functions to improve translation quality with limited parallel data.

Findings

The method achieves good translation results with partially aligned corpora.

Tiny bitexts significantly boost translation quality.

The approach is effective in low-resource translation scenarios.

Abstract

While neural machine translation (NMT) has become the new paradigm, the parameter optimization requires large-scale parallel data which is scarce in many domains and language pairs. In this paper, we address a new translation scenario in which there only exists monolingual corpora and phrase pairs. We propose a new method towards translation with partially aligned sentence pairs which are derived from the phrase pairs and monolingual corpora. To make full use of the partially aligned corpora, we adapt the conventional NMT training method in two aspects. On one hand, different generation strategies are designed for aligned and unaligned target words. On the other hand, a different objective function is designed to model the partially aligned parts. The experiments demonstrate that our method can achieve a relatively good result in such a translation scenario, and tiny bitexts can boost translation quality to a large extent.