Publication Details

Multilingual acoustic modeling for speech recognition based on Subspace Gaussian Mixture Models

BURGET Lukáš, SCHWARZ Petr, AGARWAL Mohit, AKYAZI Pinar, FENG Kai, GHOSHAL Arnab, GLEMBEK Ondřej, GOEL Nagendra K., KARAFIÁT Martin, POVEY Daniel, RASTROW Ariya, ROSE Richard and THOMAS Samuel. Multilingual acoustic modeling for speech recognition based on Subspace Gaussian Mixture Models. In: Proc. International Conference on Acoustictics, Speech, and Signal Processing. Dallas: IEEE Signal Processing Society, 2010, pp. 4334-4337. ISBN 978-1-4244-4296-6. ISSN 1520-6149.
Czech title
Multilingvální akustické modelování pro rozpoznávání řeči založené na sub-space Gaussovských modelech
Type
conference paper
Language
english
Authors
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Schwarz Petr, Ing., Ph.D. (DCGM FIT BUT)
Agarwal Mohit (IIIT)
Akyazi Pinar (UBOGAZ)
Feng Kai (HKUST)
Ghoshal Arnab (UEDIN)
Glembek Ondřej, Ing., Ph.D. (DCGM FIT BUT)
Goel Nagendra K. (GOVIVACE)
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Povey Daniel (JHU)
Rastrow Ariya (JHU)
Rose Richard (MCGILL)
Thomas Samuel (JHU)
URL
Keywords

Large vocabulary speech recognition, Subspace Gaussian mixture model, Multilingual acoustic modeling

Abstract

This paper is on a different approach to multilingual speech recognition, in which the phone sets are entirely distinct but the model has parameters not tied to specific states that are shared across languages.

Annotation

Although research has previously been done on multilingual speech recognition, it has been found to be very difficult to improve over separately trained systems. The usual approach has been to use some kind of "universal phone set" that covers multiple languages. We report experiments on a different approach to multilingual speech recognition, in which the phone sets are entirely distinct but the model has parameters not tied to specific states that are shared across languages. We use a model called a "Subspace Gaussian Mixture Model" where states' distributions are Gaussian Mixture Models with a common structure, constrained to lie in a subspace of the total parameter space. The parameters that define this subspace can be shared across languages. We obtain substantial WER improvements with this approach, especially with very small amounts of inlanguage training data.

Published
2010
Pages
4334-4337
Journal
Proc. International Conference on Acoustics, Speech, and Signal Processing, vol. 2010, no. 3, ISSN 1520-6149
Proceedings
Proc. International Conference on Acoustictics, 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{FITPUB9307,
   author = "Luk\'{a}\v{s} Burget and Petr Schwarz and Mohit Agarwal and Pinar Akyazi and Kai Feng and Arnab Ghoshal and Ond\v{r}ej Glembek and K. Nagendra Goel and Martin Karafi\'{a}t and Daniel Povey and Ariya Rastrow and Richard Rose and Samuel Thomas",
   title = "Multilingual acoustic modeling for speech recognition based on Subspace Gaussian Mixture Models",
   pages = "4334--4337",
   booktitle = "Proc. International Conference on Acoustictics, 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/9307"
}
Back to top