Edeflip: Supervised Word Translation between English and Yoruba

TL;DR

This paper explores the effectiveness of supervised embedding alignment for English-Yoruba translation, highlighting the importance of embedding quality and normalization, and addressing challenges faced by low-resource languages.

Contribution

It applies supervised embedding alignment to a low-resource language, Yoruba, revealing factors affecting translation quality and limitations of current methods.

Findings

Higher embedding quality improves translation accuracy

Normalizing embeddings increases translation precision

Low-resource languages face unique challenges in embedding alignment

Abstract

In recent years, embedding alignment has become the state-of-the-art machine translation approach, as it can yield high-quality translation without training on parallel corpora. However, existing research and application of embedding alignment mostly focus on high-resource languages with high-quality monolingual embeddings. It is unclear if and how low-resource languages may be similarly benefited. In this study, we implement an established supervised embedding alignment method for word translation from English to Yoruba, the latter a low-resource language. We found that higher embedding quality and normalizing embeddings increase word translation precision, with, additionally, an interaction effect between the two. Our results demonstrate the limitations of the state-of-the-art supervised embedding alignment when it comes to low-resource languages, for which there are additional factors that need to be taken into consideration, such as the importance of curating high-quality monolingual embeddings. We hope our work will be a starting point for further machine translation research that takes into account the challenges that low-resource languages face.