Learning robust speech representation with an articulatory-regularized variational autoencoder
This work addresses speech processing challenges for applications like denoising, though it is incremental by building on existing variational autoencoder methods.
The authors tackled the problem of improving deep generative models for speech processing by incorporating articulatory representations, resulting in faster convergence, lower reconstruction loss, and better performance in speech denoising tasks.
It is increasingly considered that human speech perception and production both rely on articulatory representations. In this paper, we investigate whether this type of representation could improve the performances of a deep generative model (here a variational autoencoder) trained to encode and decode acoustic speech features. First we develop an articulatory model able to associate articulatory parameters describing the jaw, tongue, lips and velum configurations with vocal tract shapes and spectral features. Then we incorporate these articulatory parameters into a variational autoencoder applied on spectral features by using a regularization technique that constraints part of the latent space to follow articulatory trajectories. We show that this articulatory constraint improves model training by decreasing time to convergence and reconstruction loss at convergence, and yields better performance in a speech denoising task.