Online Bayesian tree-structured transformation of HMMs with optimal model selection for speaker adaptation

IEEE Transactions on Speech and Audio Processing

Volumn 9, Issue 6, 2001, Pages 663-677

  Wang, Shaojun ^b,c   Zhao, Yunxin ^a,d  

^a IEEE

^b University of Illinois at Urbana Champaign   (United States)

^c Carnegie Mellon University ^*  (United States)

^d University of Missouri Columbia   (United States)

Author keywords

Affine transformation; Bayesian model selection; Hidden Markov models (HMMs); Linear regression (LR); Model complexity; Recursive Bayesian learning; Robust priors; Speaker adaptation; Tree structure

Indexed keywords

ADAPTIVE ALGORITHMS; COMPUTATIONAL COMPLEXITY; DYNAMIC PROGRAMMING; MATHEMATICAL MODELS; MATHEMATICAL TRANSFORMATIONS; PARAMETER ESTIMATION; PROBABILITY DENSITY FUNCTION; REGRESSION ANALYSIS; TREES (MATHEMATICS);

SPEAKER ADAPTATION;

SPEECH RECOGNITION;

EID: 0035440798     PISSN: 10636676     EISSN: None     Source Type: Journal    
DOI: 10.1109/89.943344     Document Type: Article

References (51)

1. Combined Bayesian and predictive techniques for rapid speaker adaptation of continuous density hidden Markov models
2. Tied mixture continuous parameter modeling for speech recognition
3. Comments on efficient training algorithms for HMMs using incremental estimation
4. Maximum a posteriori linear regression for hidden Markov model adaptation
5. Hybrid adaptation of tree structure and hidden Markov models for robust speech recognition
6. Online hierarchical transformation of hidden Markov models for speech recognition
7. A hybrid algorithm for speaker adaptation using MAP transformation and adaptation
8. Unsupervised hierarchical adaptation using reliable selection of cluster-dependent parameters
9. Maximum a posterior linear regression with elliptically symmetric matrix variate priors
10. The computational complexity of probabilistic inference using Bayesian belief networks
11. Predictive speaker adaptation in speech recognition
12. The sample complexity of learning fixed-structure Bayesian networks
13. Maximum likelihood estimation from incomplete data via the EM algorithm
14. Speaker adaptation using constrained estimation of Gaussian mixtures
15. Speaker adaptation using combined transformation and Bayesian methods
16. Online adaptation of hidden Markov models using incremental estimation algorithms
17. Rapid speech recognizer adaptation to new speakers
18. Mean and variance adaptation within the MLLR framework
19. Maximum likelihood linear transformations for HMM-based speech recognition
20. Maximum a posteriori estimation for multivariate Gaussian mixture observations of Markov chains
21. Efficient training algorithms for HMMs using incremental estimation
22. Semi-continuous hidden Markov models for speech signal
23. Online adaptive learning of the continuous density hidden Markov model based on approximate recursive Bayes estimate
24. Online adaptive learning of CDHMM parameters based on multiple-stream a priori evolution and posterior pooling
25. Modeling parameter dependence in speaker adaptation using multiscale tree processes
26. Schwarz, Wallace, and Rissanen: Intertwining themes in theories of model order estimation
27. A posteriori estimation of correlated jointly Gaussian mean vectors
28. On stochastic feature and model compensation approaches to robust speech recognition
29. On adaptive decision rules and decision parameter adaptation for automatic speech recognition
30. Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models
31. Generalizing case frames using a thesaurus and the MDL principle
32. Hidden Markov modeling using a dominant state sequence with application to speech recognition
33. A view of the EM algorithm that justifies incremental, sparse, and other variants
34. A Markov random field approach to Bayesian speaker adaptation
35. Speaker adaptation with autonomous model complexity control by MDL principle
36. Structural MAP speaker adaptation using hierarchical priors
37. Joint maximum a posterior adaptation of transformation and hidden Markov model parameters
38. Nonlinear compensation for stochastic matching
39. Recursive parameter estimation using incomplete data
40. Batch, incremental and instantaneous adaptation techniques for speech recognition
41. A speaker-independent continuous speech recognition system using continuous mixture Gaussian density HMM of phoneme-sized units
42. Self-learning speaker and channel adaptation based on spectral variation source decomposition
43. Probabilistic modeling with Bayesian networks for automatic speech recognition