Abstract
An estimation of parameters of a multivariate Gaussian Mixture Model is usually based on a criterion (e.g. Maximum Likelihood) that is focused mostly on training data. Therefore, testing data, which were not seen during the training procedure, may cause problems. Moreover, numerical instabilities can occur (e.g. for low-occupied Gaussians especially when working with full-covariance matrices in high-dimensional spaces). Another question concerns the number of Gaussians to be trained for a specific data set. The approach proposed in this paper can handle all these issues. It is based on an assumption that the training and testing data were generated from the same source distribution. The key part of the approach is to use a criterion based on the source distribution rather than using the training data itself. It is shown how to modify an estimation procedure in order to fit the source distribution better (despite the fact that it is unknown), and subsequently new estimation algorithm for diagonal- as well as full-covariance matrices is derived and tested.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Young, S., et al.: The HTK Book (for HTK Version 3.4), Cambridge (2006)
Diehl, F., Gales, M.J.F., Liu, X., Tomalin, M., Woodland, P.C.: Word Boundary Modelling and Full Covariance Gaussians for Arabic Speech-to-Text Systems. In: Proc. INTERSPEECH 2011, pp. 777–780 (2011)
Bell, P., King, S.: A Shrinkage Estimator for Speech Recognition with Full Covariance HMMs. In: Proc. Interspeech 2008, Brisbane, Australia (2008)
Bell, P.: Full Covariance Modelling for Speech Recognition. Ph.D. Thesis, The University of Edinburgh
Lee, Y., Lee, K.Y., Lee, J.: The Estimating Optimal Number of Gaussian Mixtures Based on Incremental k-means for Speaker Identification. International Journal of Information Technology 12(7), 13–21 (2006)
Figueiredo, M.A.T., Leitão, J.M.N., Jain, A.K.: On Fitting Mixture Models. In: Hancock, E.R., Pelillo, M. (eds.) EMMCVPR 1999. LNCS, vol. 1654, pp. 54–69. Springer, Heidelberg (1999)
Mclachlan, G.J., Peel, D.: On a Resampling Approach to Choosing the Number of Components in Normal Mixture Models. Computing Science and Statistics 28, 260–266 (1997)
Paclík, P., Novovičová, J.: Number of Components and Initialization in Gaussian Mixture Model for Pattern Recognition. In: Proc. Artificial Neural Nets and Genetic Algorithms, pp. 406–409. Springer, Wien (2001)
Schwarz, G.E.: Estimating the dimension of a model. Annals of Statistics 6(2), 461–464 (1978)
Akaike, H.: On entropy maximization principle. In: Applications of Statistics, pp. 27–41. North-Holland, Amsterdam (1977)
Machlica, L., Vanek, J., Zajic, Z.: Fast Estimation of Gaussian Mixture Model Parameters on GPU using CUDA. In: Proc. PDCAT, Gwangju, South Korea (2011)
Vanek, J., Trmal, J., Psutka, J.V., Psutka, J.: Optimized Acoustic Likelihoods Computation for NVIDIA and ATI/AMD Graphics Processors. IEEE Transactions on Audio, Speech and Language Processing 20(6), 1818–1828 (2012)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer International Publishing Switzerland
About this paper
Cite this paper
Vaněk, J., Machlica, L., Psutka, J.V., Psutka, J. (2013). Covariance Matrix Enhancement Approach to Train Robust Gaussian Mixture Models of Speech Data. In: Železný, M., Habernal, I., Ronzhin, A. (eds) Speech and Computer. SPECOM 2013. Lecture Notes in Computer Science(), vol 8113. Springer, Cham. https://doi.org/10.1007/978-3-319-01931-4_13
Download citation
DOI: https://doi.org/10.1007/978-3-319-01931-4_13
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-01930-7
Online ISBN: 978-3-319-01931-4
eBook Packages: Computer ScienceComputer Science (R0)