Phone-ing it in: Towards Flexible Multi-Modal Language Model Training by Phonetic Representations of Data

Anonymous

Abstract: Multi-modal techniques offer significant untapped potential to unlock improved NLP functionality for local languages. However, many advances in language model pre-training are focused on text, a fact that only increases systematic inequalities in the performance of NLP tasks across the world's languages. In this work, we propose a multi-modal approach to train language models using whatever text and/or audio data might be available in a language. Initial experiments using Swahili and Kinyarwanda data suggest the viability of the approach for downstream Named Entity Recognition (NER) tasks, with models pre-trained on phone data showing an improvement of up to 6\% F1-score above models that are trained from scratch.