Africa-Centric Self-Supervised Pretraining for Multilingual Speech Representation in a Sub-Saharan Context

ICLR 2024 Workshop AfricaNLP Submission21 Authors

Keywords: Self-supervised speech representation, African languages, Multilingual ASR, HuBERT

TL;DR: First publicly shared Self-supervised multilingual speech model trained exclusively on sub-Saharan spoken data

Abstract:

We present the first self-supervised multilingual speech model trained exclusively on African speech.
The model learned from nearly 60 000 hours of unlabeled speech segments in 21 languages and dialects spoken in sub-Saharan Africa.
On the SSA subset of the FLEURS-102 dataset, our approach based on a HuBERT$_{base}$ (0.09B) architecture shows competitive results, for ASR downstream task, compared to the w2v-bert-51 (0.6B) pre-trained model proposed in the FLEURS benchmark, while being more efficient by using 7x less data and 6x less parameters.
Furthermore, in the context of a LID downstream task, our approach outperforms FLEURS baselines accuracy by over 22\%.

Submission Number: 21