Selecting Artificially-Generated Sentences for Fine-Tuning Neural Machine Translation

TL;DR

This paper investigates how artificially-generated sentence pairs, selected using data-selection algorithms, can enhance German-to-English neural machine translation models beyond traditional training with only authentic data.

Contribution

It demonstrates that artificially-generated sentences, when combined with data-selection algorithms, can outperform authentic data in training NMT models.

Findings

Artificially-generated sentences can be more beneficial than authentic pairs.

Combining artificial data with data-selection algorithms improves translation performance.

Small, domain-specific artificial datasets can boost NMT accuracy.

Abstract

Neural Machine Translation (NMT) models tend to achieve best performance when larger sets of parallel sentences are provided for training. For this reason, augmenting the training set with artificially-generated sentence pairs can boost performance. Nonetheless, the performance can also be improved with a small number of sentences if they are in the same domain as the test set. Accordingly, we want to explore the use of artificially-generated sentences along with data-selection algorithms to improve German-to-English NMT models trained solely with authentic data. In this work, we show how artificially-generated sentences can be more beneficial than authentic pairs, and demonstrate their advantages when used in combination with data-selection algorithms.