Discriminative training of HMM using maximum normalized likelihood algorithm

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

Volumn 1, Issue , 2001, Pages 497-500

  Markov, Konstuntin ^a   Nakagawa, Seiichi ^a   Nakamura, Satoshi ^b  

^a TOYOHASHI UNIVERSITY OF TECHNOLOGY   (Japan)

^b ADVANCED TELECOMMUNICATIONS RESEARCH INSTITUTE INTERNATIONAL   (Japan)

Author keywords

[No Author keywords available]

Indexed keywords

DISCRIMINATIVE TRAINING; EXPECTATION MAXIMIZATION; MAXIMUM MUTUAL INFORMATION; MAXIMUM NORMALIZED LIKELIHOOD ALGORITHM; MINIMUM CLASSIFICATION ERROR;

MATHEMATICAL MODELS; MAXIMUM LIKELIHOOD ESTIMATION; OPTIMIZATION; PARAMETER ESTIMATION; PROBABILITY;

SPEECH RECOGNITION;

EID: 0034841731     PISSN: 15206149     EISSN: None     Source Type: Journal    
DOI: 10.1109/ICASSP.2001.940876     Document Type: Article

References (10)

1. Prototype-based minimum classification error/generalized probabilistic descent training for various speech units
2. Minimum classification error rate methods for speech recognition
3. High-performance connected digit recognition using maximum mutual information estimation
4. MMIE training of large vocabulary recognition systems
5. Discriminative training of GMM for speaker identification
6. Pattern recognition using a family of design algorithms based upon the generalized probabilistic descent method
7. Frame discrimination training of HMMs for large vocabulary speech recognition
8. Discriminative training of GMM using a modified EM algorithm for speaker recognition
9. Text-independent speaker recognition based on frame level likelihood transformations
10. Evaluation of segmental unit input HMM