Benchmarking Generative Latent Variable Models for Speech
Abstract
Stochastic latent variable models (LVMs) achieve state-of-the-art performance on natural image generation but are still inferior to deterministic models on speech. In this paper, we develop a speech benchmark of popular temporal LVMs and compare them against state-of-the-art deterministic models. We report the likelihood, which is a much used metric in the image domain, but rarely, or incomparably, reported for speech models. To assess the quality of the learned representations, we also compare their usefulness for phoneme recognition. Finally, we adapt the Clockwork VAE, a state-of-the-art temporal LVM for video generation, to the speech domain. Despite being autoregressive only in latent space, we find that the Clockwork VAE can outperform previous LVMs and reduce the gap to deterministic models by using a hierarchy of latent variables.
Cite
@article{arxiv.2202.12707,
title = {Benchmarking Generative Latent Variable Models for Speech},
author = {Jakob D. Havtorn and Lasse Borgholt and Søren Hauberg and Jes Frellsen and Lars Maaløe},
journal= {arXiv preprint arXiv:2202.12707},
year = {2022}
}
Comments
Accepted at the 2022 ICLR workshop on Deep Generative Models for Highly Structured Data (https://deep-gen-struct.github.io)