Publication Details
Bayesian phonotactic language model for acoustic unit discovery
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Černocký Jan, prof. Dr. Ing. (DCGM)
Kesiraju Santosh, Ph.D. (DCGM)
Bayesian non-parametric, Variational Bayes, acoustic unit discovery
Recent work on Acoustic Unit Discovery (AUD) has led to the development of a non-parametric Bayesian phone-loop model where the prior over the probability of the phone-like units is assumed to be sampled from a Dirichlet Process (DP). In this work, we propose to improve this model by incorporating a Hierarchical Pitman-Yor based bigram Language Model on top of the units transitions. This new model makes use of the phonotactic context information but assumes a fixed number of units. To remedy this limitation we first train a DP phoneloop model to infer the number of units, then, the bigram phone-loop is initialized from the DP phone-loop and trained until convergence of its parameters. Results show an absolute improvement of 1-2%on the Normalized Mutual Information (NMI) metric. Furthermore, we show that, combined with Multilingual Bottleneck (MBN) features the model yields a same or higher NMI as an English phone recogniser trained on TIMIT.
This article is about Bayesian phonotactic language model for acoustic unit discovery (AUD), which has led to the development of a non-parametric Bayesian phone-loop model
@inproceedings{BUT144452,
author="Lucas Antoine Francois {Ondel} and Lukáš {Burget} and Jan {Černocký} and Santosh {Kesiraju}",
title="Bayesian phonotactic language model for acoustic unit discovery",
booktitle="Proceedings of ICASSP 2017",
year="2017",
pages="5750--5754",
publisher="IEEE Signal Processing Society",
address="New Orleans",
doi="10.1109/ICASSP.2017.7953258",
isbn="978-1-5090-4117-6",
url="https://www.fit.vut.cz/research/publication/11472/"
}