Improving Machine Translation with Phrase Pair Injection and Corpus Filtering

TL;DR

This paper introduces a method combining phrase pair injection and corpus filtering to enhance neural machine translation, achieving significant BLEU score improvements for low-resource language pairs.

Contribution

It proposes a novel approach that extracts and injects phrase pairs and filters corpora to improve translation quality in low-resource settings.

Findings

Up to 2.7 BLEU point improvements on FLORES data

Effective for 3 low-resource language pairs and 6 translation directions

Enhances NMT performance using pseudo-parallel corpus augmentation

Abstract

In this paper, we show that the combination of Phrase Pair Injection and Corpus Filtering boosts the performance of Neural Machine Translation (NMT) systems. We extract parallel phrases and sentences from the pseudo-parallel corpus and augment it with the parallel corpus to train the NMT models. With the proposed approach, we observe an improvement in the Machine Translation (MT) system for 3 low-resource language pairs, Hindi-Marathi, English-Marathi, and English-Pashto, and 6 translation directions by up to 2.7 BLEU points, on the FLORES test data. These BLEU score improvements are over the models trained using the whole pseudo-parallel corpus augmented with the parallel corpus.