Bayesian learning for hidden Markov model with Gaussian mixture state observation densities
作者： Jean-Luc Gauvain , Chin-Hui Lee
DOI: 10.1016/0167-6393(92)90015-Y
关键词:
摘要: Abstract An investigation into the use of Bayesian learning parameters a multivariate Gaussian mixture density has been carried out. In framework continuous hidden Markov model (CDHMM), serves as unified approach for parameter smoothing, speaker adaptation, clustering and corrective training. The goal is to enhance robustness in CDHMM-based speech recognition system so improve performance. Our incorporate prior knowledge training process form densities HMM parameters. theoretical basis this procedure presented results applying it are given.
isca-speech.org
elsevier.com参考文章(14)
Marco Ferretti, Stefano Scarci, Large-vocabulary speech recognition with speaker-adapted codebook and HMM parameters. conference of the international speech communication association. pp. 2154- 2156 ,(1989)
F. Jelinek, Interpolated estimation of Markov source parameters from sparse data Proc. Workshop on Pattern Recognition in Practice, 1980. pp. 381- 397 ,(1980)
R. Zelinski, F. Class, A learning procedure for speaker-dependent word recognition systems based on sequential processing of input tokens international conference on acoustics, speech, and signal processing. ,vol. 8, pp. 1053- 1056 ,(1983) , 10.1109/ICASSP.1983.1171906
P. Brown, Chin-Hui Lee, J. Spohrer, Bayesian adaptation in speech recognition international conference on acoustics, speech, and signal processing. ,vol. 8, pp. 761- 764 ,(1983) , 10.1109/ICASSP.1983.1172084
Jean-Luc Gauvain, Chin-Hui Lee, Bayesian learning of Gaussian mixture densities for hidden Markov models human language technology. pp. 272- 277 ,(1991) , 10.3115/112405.112457
Morris Herman DeGroot, Optimal Statistical Decisions ,(1970)
C.-H. Lee, E. Giachin, L. R. Rabiner, R. Pieraccini, A. E. Rosenberg, Improved acoustic modeling for continuous speech recognition human language technology. pp. 319- 326 ,(1990) , 10.3115/116580.116686
Y. Normandin, S.D. Morgera, An improved MMIE training algorithm for speaker-independent, small vocabulary, continuous speech recognition international conference on acoustics, speech, and signal processing. pp. 537- 540 ,(1991) , 10.1109/ICASSP.1991.150395
A. P. Dempster, N. M. Laird, D. B. Rubin, Maximum Likelihood from Incomplete Data Via theEMAlgorithm Journal of the Royal Statistical Society: Series B (Methodological). ,vol. 39, pp. 1- 22 ,(1977) , 10.1111/J.2517-6161.1977.TB01600.X
Xuedong Huang, Fil Alleva, Satoru Hayamizu, Hsiao-Wuen Hon, Mei-Yuh Hwang, Kai-Fu Lee, Improved hidden Markov modeling for speaker-independent continuous speech recognition human language technology. pp. 327- 331 ,(1990) , 10.3115/116580.116687