Publication Details

A novel estimation of feature-space MLLR for full-covariance models

GHOSHAL Arnab, POVEY Daniel, AGARWAL Mohit, AKYAZI Pinar, BURGET Lukáš, FENG Kai, GLEMBEK Ondřej, GOEL Nagendra K., KARAFIÁT Martin, RASTROW Ariya, ROSE Richard, SCHWARZ Petr and THOMAS Samuel. A novel estimation of feature-space MLLR for full_covariance models. In: Proc. International Conference on Acoustics, Speech, and Signal Processing. Dallas: IEEE Signal Processing Society, 2010, pp. 4310-4313. ISBN 978-1-4244-4296-6. ISSN 1520-6149.
Czech title
Inovovaný odhad MLLR v prostoru parametrů pro plně kovarianční modely
Type
conference paper
Language
english
Authors
Ghoshal Arnab (UEDIN)
Povey Daniel (JHU)
Agarwal Mohit (IIIT)
Akyazi Pinar (UBOGAZ)
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Feng Kai (HKUST)
Glembek Ondřej, Ing., Ph.D. (DCGM FIT BUT)
Goel Nagendra K. (GOVIVACE)
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Rastrow Ariya (JHU)
Rose Richard (MCGILL)
Schwarz Petr, Ing., Ph.D. (DCGM FIT BUT)
Thomas Samuel (JHU)
URL
Keywords

Speech recognition, Speaker adaptation, Hidden Markov models, Optimization methods, Linear algebra

Abstract

The paper is on a novel estimation of feature-space MLLR for full-covariance models. We present a new approach for full-covariance Gaussian models.

Annotation

In this paper we present a novel approach for estimating featurespace maximum likelihood linear regression (fMLLR) transforms for full-covariance Gaussian models by directly maximizing the likelihood function by repeated line search in the direction of the gradient. We do this in a pre-transformed parameter space such that an approximation to the expected Hessian is proportional to the unit matrix. The proposed algorithm is as efficient or more efficient than standard approaches, and is more flexible because it can naturally be combined with sets of basis transforms and with full covariance and subspace precision and mean (SPAM) models.

Published
2010
Pages
4310-4313
Journal
Proc. International Conference on Acoustics, Speech, and Signal Processing, vol. 2010, no. 3, ISSN 1520-6149
Proceedings
Proc. International Conference on Acoustics, Speech, and Signal Processing
Conference
International Conference on Acoustics, Speech, and Signal Processing 2010, Dallas, US
ISBN
978-1-4244-4296-6
Publisher
IEEE Signal Processing Society
Place
Dallas, US
BibTeX
@INPROCEEDINGS{FITPUB9308,
   author = "Arnab Ghoshal and Daniel Povey and Mohit Agarwal and Pinar Akyazi and Luk\'{a}\v{s} Burget and Kai Feng and Ond\v{r}ej Glembek and K. Nagendra Goel and Martin Karafi\'{a}t and Ariya Rastrow and Richard Rose and Petr Schwarz and Samuel Thomas",
   title = "A novel estimation of feature-space MLLR for full-covariance models",
   pages = "4310--4313",
   booktitle = "Proc. International Conference on Acoustics, Speech, and Signal Processing",
   journal = "Proc. International Conference on Acoustics, Speech, and Signal Processing",
   volume = 2010,
   number = 3,
   year = 2010,
   location = "Dallas, US",
   publisher = "IEEE Signal Processing Society",
   ISBN = "978-1-4244-4296-6",
   ISSN = "1520-6149",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/9308"
}
Back to top