Bayesian Models for Unit Discovery on a Very Low Resource Language

TL;DR

This paper demonstrates the effectiveness of Bayesian models for unsupervised acoustic unit discovery and word segmentation in a low-resource language, outperforming baseline methods and integrating multilingual information.

Contribution

It applies state-of-the-art Bayesian models to real low-resource language data and shows how informative priors from other languages improve unit discovery and segmentation.

Findings

Bayesian models outperform Segmental-DTW baseline in word segmentation.

Incorporating multilingual priors improves acoustic unit discovery.

Bayesian approaches are effective in real low-resource language scenarios.

Abstract

Developing speech technologies for low-resource languages has become a very active research field over the last decade. Among others, Bayesian models have shown some promising results on artificial examples but still lack of in situ experiments. Our work applies state-of-the-art Bayesian models to unsupervised Acoustic Unit Discovery (AUD) in a real low-resource language scenario. We also show that Bayesian models can naturally integrate information from other resourceful languages by means of informative prior leading to more consistent discovered units. Finally, discovered acoustic units are used, either as the 1-best sequence or as a lattice, to perform word segmentation. Word segmentation results show that this Bayesian approach clearly outperforms a Segmental-DTW baseline on the same corpus.