Zero-Shot Language Transfer vs Iterative Back Translation for Unsupervised Machine Translation

TL;DR

This paper compares zero-shot transfer learning and unsupervised machine translation for low-resource languages, analyzing how data size and domain influence performance, with experiments available on GitHub.

Contribution

It provides a comparative analysis of zero-shot transfer and unsupervised MT, highlighting the effects of data size and domain on their effectiveness.

Findings

Data size significantly impacts both methods' performance.

Domain differences affect unsupervised MT results.

Code for experiments is publicly available on GitHub.

Abstract

This work focuses on comparing different solutions for machine translation on low resource language pairs, namely, with zero-shot transfer learning and unsupervised machine translation. We discuss how the data size affects the performance of both unsupervised MT and transfer learning. Additionally we also look at how the domain of the data affects the result of unsupervised MT. The code to all the experiments performed in this project are accessible on Github.