Probing phoneme, language and speaker information in unsupervised speech representations

TL;DR

This paper investigates the types of linguistic and speaker information encoded in unsupervised CPC speech representations, revealing that bilingual models better discriminate languages and that gender information increases with more clusters, with some trade-offs in phoneme discrimination.

Contribution

It provides a comprehensive analysis of phoneme, language, and gender information in CPC speech representations, highlighting differences between monolingual and bilingual models and their downstream task implications.

Findings

Gender and phone class information are present in both models.

Language information is more salient in bilingual models.

More clusters encode more gender information.

Abstract

Unsupervised models of representations based on Contrastive Predictive Coding (CPC)[1] are primarily used in spoken language modelling in that they encode phonetic information. In this study, we ask what other types of information are present in CPC speech representations. We focus on three categories: phone class, gender and language, and compare monolingual and bilingual models. Using qualitative and quantitative tools, we find that both gender and phone class information are present in both types of models. Language information, however, is very salient in the bilingual model only, suggesting CPC models learn to discriminate languages when trained on multiple languages. Some language information can also be retrieved from monolingual models, but it is more diffused across all features. These patterns hold when analyses are carried on the discrete units from a downstream clustering model. However, although there is no effect of the number of target clusters on phone class and language information, more gender information is encoded with more clusters. Finally, we find that there is some cost to being exposed to two languages on a downstream phoneme discrimination task.