High-Quality Data Augmentation for Low-Resource NMT: Combining a Translation Memory, a GAN Generator, and Filtering

TL;DR

This paper introduces a novel data augmentation method for low-resource neural machine translation by combining a translation memory, a GAN generator, and a filtering process to improve translation quality.

Contribution

It presents a new approach that integrates a GAN, translation memory, and filtering to enhance data quality and quantity in low-resource NMT settings.

Findings

Improved translation quality with augmented data.

Effective filtering reduces low-quality synthetic data.

GAN-based augmentation outperforms traditional methods.

Abstract

Back translation, as a technique for extending a dataset, is widely used by researchers in low-resource language translation tasks. It typically translates from the target to the source language to ensure high-quality translation results. This paper proposes a novel way of utilizing a monolingual corpus on the source side to assist Neural Machine Translation (NMT) in low-resource settings. We realize this concept by employing a Generative Adversarial Network (GAN), which augments the training data for the discriminator while mitigating the interference of low-quality synthetic monolingual translations with the generator. Additionally, this paper integrates Translation Memory (TM) with NMT, increasing the amount of data available to the generator. Moreover, we propose a novel procedure to filter the synthetic sentence pairs during the augmentation process, ensuring the high quality of the data.