Simultaneous Multi-Pivot Neural Machine Translation

TL;DR

This paper introduces a multi-pivot approach for simultaneous neural machine translation, improving translation quality in low-resource scenarios by leveraging multiple pivot languages concurrently.

Contribution

It proposes a novel multi-pivot translation method that enhances simultaneous NMT performance by translating through multiple pivots simultaneously.

Findings

Using two pivot languages improves BLEU scores by up to 5.8.

Multi-pivot translation reduces quality deterioration in low-resource, real-time settings.

The approach demonstrates effectiveness on Arabic-English translation via French and Spanish.

Abstract

Parallel corpora are indispensable for training neural machine translation (NMT) models, and parallel corpora for most language pairs do not exist or are scarce. In such cases, pivot language NMT can be helpful where a pivot language is used such that there exist parallel corpora between the source and pivot and pivot and target languages. Naturally, the quality of pivot language translation is more inferior to what could be achieved with a direct parallel corpus of a reasonable size for that pair. In a real-time simultaneous translation setting, the quality of pivot language translation deteriorates even further given that the model has to output translations the moment a few source words become available. To solve this issue, we propose multi-pivot translation and apply it to a simultaneous translation setting involving pivot languages. Our approach involves simultaneously translating a source language into multiple pivots, which are then simultaneously translated together into the target language by leveraging multi-source NMT. Our experiments in a low-resource setting using the N-way parallel UN corpus for Arabic to English NMT via French and Spanish as pivots reveals that in a simultaneous pivot NMT setting, using two pivot languages can lead to an improvement of up to 5.8 BLEU.