Parameter learning in a hidden trajectory model

Details Parameter learning in a hidden trajectory model Parameter learning in a hidden trajectory model

2011-08-30

2011-08-30

Dorvil, Richemond (Department: 2626)

Data processing: speech signal processing, linguistics, language

Speech signal processing

Recognition

C704S218000, C704S244000

Reexamination Certificate

active

08010356

ABSTRACT:
Parameters for distributions of a hidden trajectory model including means and variances are estimated using an acoustic likelihood function for observation vectors as an objection function for optimization. The estimation includes only acoustic data and not any intermediate estimate on hidden dynamic variables. Gradient ascent methods can be developed for optimizing the acoustic likelihood function.

REFERENCES:
patent: 6076058 (2000-06-01), Chengalvarayan
patent: 6609093 (2003-08-01), Gopinath et al.
patent: 6618699 (2003-09-01), Lee et al.
patent: 7010167 (2006-03-01), Ordowski et al.
patent: 2003/0216911 (2003-11-01), Deng et al.
patent: 2003/0225719 (2003-12-01), Juang et al.
patent: 2004/0019483 (2004-01-01), Deng et al.
patent: 2004/0078198 (2004-04-01), Hernandez-Abrego et al.
patent: 2004/0143435 (2004-07-01), Deng et al.
patent: 2004/0199382 (2004-10-01), Bazzi et al.
patent: 2004/0199386 (2004-10-01), Attias et al.
patent: 2004/0260548 (2004-12-01), Attias et al.
patent: 2006/0178887 (2006-08-01), Webber
patent: 2006/0229875 (2006-10-01), Acero et al.
patent: 2007/0129943 (2007-06-01), Lei et al.
J-L. Zhou, F. Seide, and L. Deng. Coarticulation modeling by embedding a target-directed hidden trajectory model into HMM-model and training. In Proceedings ICASSP, vol. 1, pp. 744-747, 2003.
L. Xu and M. I. Jordan. On convergence properties of the EM algorithm for Gaussian mixtures. Neural Computation, 8:129-151, 1996.
K. Tokuda, H. Zen, and T. Kitamura, “Reformulating the HMM as a trajectory model,” in Proceedings Beyond HMM Workshop, Tokyo, Dec. 2004.
L. Deng, L. J. Lee, H. Attias, and A. Acero. “A structured speech model with continuous hidden dynamics and prediction-residual training for tracking vocal tract resonances,” Proc. ICASSP, vol. 1, 2004, pp. 557-560.
Schraudolf, N.: Gradient-based manipulation of non-parametric entropy estimates. IEEE Trans. on Neural Networks 16 (2004) 159-195.
L. Deng, I. Bazzi, and A. Acero, “Tracking vocal tract resonances using an analytical nonlinear predictor and a target-guided temporal constraint,” in Proc. EUROSPEECH, 2003, pp. 73-76.
Dusan, S., 2000. Statistical estimation of articulatory trajectories from the speech signal using dynamical and phonological constraints. Ph.D. Thesis, Department of Electrical and Computer Engineering, University of Waterloo, Waterloo, Canada, April.
Rigdon. “Not Positive Definite Matrices—Causes and Cures” 1997. Retrieved from: http://www2.gsu.edu˜mkteer
pdmatri.html Sep. 3, 2010.
Deng et al. “Learning Statistically Characterized Resonance Targets in a Hidden Trajectory Model of Speech Coarticulation and Reduction” Sep. 4-8, 2005.
Deng et al. “A Long-Contextual-Span Model of Resonance Dynamics for Speech Recognition: Parameter Learning and Recognizer Evaluation” Nov. 28, 2005.
Yu et al. “Evaluation of a Long-Contextual-Span Hidden Trajectory Model and Phonetic Recognizer Using A Lattice Search” Sep. 4-8, 2005.
Deng et al. “A Hidden Trajectory Modelwith Bi-Directional Target-Filtering: Cascaded vs. Integrated Implementation for Phonetic Recognition” 2005.
Deng et al. “A Generative Modeling Framework for Structured Hidden Speech Dynamics” 2005.
Deng et al. “A Structured Speech Modelwith Continuous Hidden Dynamics and Prediction-Residual Training for Tracking Vocal Tract Resonances” 2004.
Pinheiro et al. “Unconstrained parametrizations for variance-covariance matrices” 1996.
Moller et al. “A Scaled Conjugate Gradient Algorithm for Fast Supervised Learning” 1993.
Deng et al. “Novel Acoustic Modeling With Structured Hidden Dynamics for Speech Coarticulation and Reduction” 2004.
Deng et al. “A Bidirectional Target-Filtering Model of Speech Coarticulation and Reduction: Two-Stage Implementation for Phonetic Recognition” Jan. 2006.
U.S. Appl. No. 10/944,262, filed Sep. 14, 2004, Deng et al.
U.S. Appl. No. 11/069,474, filed Mar. 1, 2005, Acero et al.
U.S. Appl. No. 11/093,833, filed Mar. 30, 2005, Acero et al.
U.S. Appl. No. 11/356,905, filed Feb. 17, 2006, Li et al.
L. Deng, D. Yu, and A. Acero. “A quantitative model for formant dynamics and contextually assimilated reduction in fluent speech”, in Proc. ICSLP, pp. 719-722, Jeju, Korea, 2004.
L. Deng, X. Li, D. Yu, and A. Acero. “A hidden trajectory model with bi-directional target-filtering: Cascaded vs. integrated implementation for phonetic recognition”, in Proc. IEEE ICASSP, pp. 337-340, Mar. 2005, Philadelphia.
L. Deng, D. Yu, and A. Acero. “Learning statistically characterized resonance targets in a hidden trajectory model of speech coarticulation and reduction”, in Proc. Interspeech 2005, Lisbon, Sep. 2005, pp. 1097-1100.
D. Yu, L. Deng and A. Acero. “Evaluation of a long-contextual-span trajectory model and phonetic recognizer using A* lattice search”, in Proc. Interspeech, Lisbon, Sep. 2005, pp. 553-556.
M. Akagi. “Modeling of contextual effects based on spectral peak interaction,” in J. Acoust. Soc. Am., vol. 93, No. 2, pp. 1076-1086, 1993.
J. Glass. “A probablistic framework for segment-based speech recognition,” in Computer Speech and Language, vol. 17, 2003, pp. 137-152.
J. Krause and L. Braida. “Acoustic properties of naturally produced clear speech at normal speaking rates,” in J. Acoust. Soc. Am., vol. 115, No. 1, pp. 362-378, 2004.

Affiliated with

Also associated with

Rating

Parameter learning in a hidden trajectory model does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Parameter learning in a hidden trajectory model, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Parameter learning in a hidden trajectory model will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2752537