Improved acoustic-to-articulatory inversion using representations from pretrained self-supervised learning models

TL;DR

This paper explores the use of pretrained self-supervised learning features for acoustic-to-articulatory inversion, showing they perform comparably to traditional MFCC features across various model complexities and configurations.

Contribution

It demonstrates that SSL features like TERA and DeCoAR are effective for AAI, offering a viable alternative to traditional acoustic features across different neural network models.

Findings

SSL features achieve high correlation scores close to MFCC in AAI tasks.

Performance of SSL features is consistent across different model sizes.

SSL features work well in subject-specific, pooled, and fine-tuned configurations.

Abstract

In this work, we investigate the effectiveness of pretrained Self-Supervised Learning (SSL) features for learning the mapping for acoustic to articulatory inversion (AAI). Signal processing-based acoustic features such as MFCCs have been predominantly used for the AAI task with deep neural networks. With SSL features working well for various other speech tasks such as speech recognition, emotion classification, etc., we experiment with its efficacy for AAI. We train on SSL features with transformer neural networks-based AAI models of 3 different model complexities and compare its performance with MFCCs in subject-specific (SS), pooled and fine-tuned (FT) configurations with data from 10 subjects, and evaluate with correlation coefficient (CC) score on the unseen sentence test set. We find that acoustic feature reconstruction objective-based SSL features such as TERA and DeCoAR work well for AAI, with SS CCs of these SSL features reaching close to the best FT CCs of MFCC. We also find the results consistent across different model sizes.