Robust speaker adaptation by weighted model averaging based on the minimum description length criterion

IEEE Transactions on Audio, Speech and Language Processing

Volumn 15, Issue 2, 2007, Pages 652-660

  Cui, Xiaodong ^a,b   Alwan, Abeer ^a  

^a UNIVERSITY OF CALIFORNIA   (United States)

^b IBM T J WATSON RESEARCH CENTER   (United States)

Author keywords

Maximum likelihood linear regression (MLLR); Minimum description length (MDL); Model averaging; Speaker adaptation

Indexed keywords

COMPUTATIONAL ADVANTAGES; CONSISTENT PERFORMANCE; MAXIMUM LIKELIHOOD LINEAR REGRESSION (MLLR); MAXIMUM-LIKELIHOOD ESTIMATIONS; MINIMUM DESCRIPTION LENGTH (MDL); MINIMUM DESCRIPTION LENGTH CRITERION; MINIMUM DESCRIPTION LENGTH PRINCIPLES; MODEL AVERAGING; REGRESSION TREES; SPEAKER ADAPTATION; WEIGHTED MODELS;

BLOCK CODES; MODEL STRUCTURES; PARAMETER ESTIMATION; REGRESSION ANALYSIS; TREES (MATHEMATICS);

MAXIMUM LIKELIHOOD ESTIMATION;

EID: 64149094039     PISSN: 15587916     EISSN: None     Source Type: Journal    
DOI: 10.1109/TASL.2006.876773     Document Type: Article

References (29)

1. C. Leggetter and P.Woodland, "Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models," Comput. Speech Lang., vol. 9, pp. 171-185, 1995.
2. M. Padmanabhan and S. Dharanipragada, "Maximum-likelihood nonlinear transformation for acoustic adaptation," IEEE Trans. Speech Audio Process., vol. 12, no. 6, pp. 572-578, Nov. 2004.
3. J.-L. Gauvain and C.-H. Lee, "Maximum a posteriori estimation for multivariate Gaussian mixture observations of Markov chains," IEEE Trans. Speech Audio Process., vol. 2, no. 2, pp. 291-298, Mar. 1994.
4. K. Shinoda and C. Lee, "Structural MAP speaker adaptation using hierarchical priors," in Proc. IEEE-SPWorkshop Automatic Speech Recognition and Understanding, 1997, pp. 381-388.
5. K. Shinoda and C.-H. Lee, "A structural Bayes approach to speaker adaptation," IEEE Trans. Speech Audio Process., vol. 9, no. 3, pp. 276-287, May 2001.
6. V. Digalakis et al., "Rapid speech recognizer adaptation to new speakers," in Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, 1999, pp. 765-768.
7. E. Bocchieri et al., "Correlation modeling of MLLR transform biases for rapid HMM adaptation to new speakers," in Proc. IEEE Int. Conf. Acoustics, Speech Signal Processing, 1999, pp. 773-776.
8. R. Kuhn, J. Junqua, P. Nguyen, and N. Niedzielski, "Rapid speaker adaptaion in cigenvoice space," IEEE Trans. Speech Audio Process., vol. 8, no. 6, pp. 695-707, Nov. 2000.
9. S. D. Peters, "Hypothesis-driven adaptation (HYDRA: a flexible eigenvoice architecture)," in Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, 2001, pp. 349-352.
10. Y. Tsao et al., "Segmental eigenvoice for rapid speaker adaptation," in Proc. Eur. Conf. Speech Communication and Technology, 2001, pp. 1269-1272.
11. B. Mak et al., "A study of various composite kernels for kernel eigenvoice speaker adaptation," in Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, 2004, pp. 325-328.
12. J. Rissanen, "Modeling by shortest data description," Automatica, vol. 14, pp. 465-471, 1978.
13. K. Shinoda and T. Watanabe, "Speaker adaptation with autonomous model complexity control by MDL principle," in Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, 1996, pp. 717-720.
14. I. Illina, "Tree-structured maximum a posteriori adaptation for a segment-based speech recognition system," in Proc. Int. Conf. Spoken Language Process., 2002.
15. M. Hansen and B. Yu, "Model selectin and the principle of minimum description length," J. Amer. Statist. Assoc., vol. 96, no. 454, pp. 746-774, 2001.
16. A. Barron, J. Rissanen, and B. Yu, "The minimum description length principle in coding and modeling," IEEE Trans. Inf. Theory, vol. 44, no. 6, pp. 2743-2760, Oct. 1998.
17. R. Stine, Model Selection Using Information Theory and the MDL Principle [Online]. Available: http://www-stat.wharton.upenn.edu/̃stine
18. K. Shinoda and T. Watanabe, "Acoustic modeling based on the MDL principle for speech recognition," in Proc. Eur. Conf. Speech Communication and Technology, 1997, pp. 99-102.
19. S.Wang and Y. Zhao, "Online Bayesian tree-structured transformation of HMMs with optimal model selection for speaker adaptation," IEEE Trans. Speech Audio Process., vol. 9, no. 6, pp. 663-677, Nov. 2001.
20. S.Young et al., The HTK Book (Version 3.1). Cambridge, U.K.: Cambridge Univ. Eng. Dept., 2001.
21. L. Rabiner and B.-H. Juang, Fundamentals of Speech Recognition. Englewood Cliffs, NJ: Prentice Hall, 1993.
22. P. Elias, "Universal codeword sets and representation of the intergers," IEEE Trans. Inf. Theory, vol. IT-21, no. 2, pp. 194-203, Mar. 1975.
23. J. Rissanen, "A universal prior for integers and estimation by minimum description length," Ann. Statist., vol. 11, pp. 416-431, 1983.
24. M. Pitz and H. Ney, "Vocal tract normalization as linear transformation of MFCC," in Proc. Eur. Conf. Speech Communication and Technology, 2003, pp. 1445-1448.
25. G. Ding, Y. Zhu, C. Li, and B. Xu, "Implementing vocal tract length normalization in the MLLR framework," in Proc. Int. Conf. Spoken Language Processing, 2002, pp. 1389-1392.
26. X. Cui and A. Alwan, "Adaptation of children's speech with limited data based on formant-like peak alignment," Comput. Speech Language, to be published.
27. L. Gillick and S. Cox, "Some statistical issues in the comparison of speech recognition algorithms," in Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, 1989, pp. 532-535.
28. H. Akaike, "A new look at the statistical model identification, " IEEE Trans. Autom. Control, vol. 19, no. 6, pp. 716-723, Dec. 1974.
29. G. Schwartz, "Estimating the dimension of a model," Ann. Statist., vol. 6, pp. 461-464, 1978.