Low-Resource Cross-Lingual Adaptive Training for Nigerian Pidgin

TL;DR

This paper introduces a cross-lingual adaptive training framework for Nigerian Pidgin, leveraging English pre-trained models and data augmentation to improve language understanding and translation in low-resource settings.

Contribution

It proposes a novel adaptive training framework combining continual and task adaptive training, and demonstrates its effectiveness on Nigerian Pidgin tasks using a large-scale parallel corpus.

Findings

English pre-trained models outperform multilingual models on Pidgin tasks

Data augmentation and back-translation significantly improve performance

Up to 2.38 BLEU score improvements achieved

Abstract

Developing effective spoken language processing systems for low-resource languages poses several challenges due to the lack of parallel data and limited resources for fine-tuning models. In this work, we target on improving upon both text classification and translation of Nigerian Pidgin (Naija) by collecting a large-scale parallel English-Pidgin corpus and further propose a framework of cross-lingual adaptive training that includes both continual and task adaptive training so as to adapt a base pre-trained model to low-resource languages. Our studies show that English pre-trained language models serve as a stronger prior than multilingual language models on English-Pidgin tasks with up to 2.38 BLEU improvements; and demonstrate that augmenting orthographic data and using task adaptive training with back-translation can have a significant impact on model performance.