Minimum phone error and I-smoothing for improved discriminative training

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

Volumn 1, Issue , 2002, Pages

  Povey, D ^a   Woodland, P C ^a  

^a UNIVERSITY OF CAMBRIDGE   (United Kingdom)

Author keywords

[No Author keywords available]

Indexed keywords

BIT ERROR RATE; ERROR ANALYSIS; INTERPOLATION; LEARNING SYSTEMS; MARKOV PROCESSES; MATHEMATICAL MODELS; MAXIMUM LIKELIHOOD ESTIMATION; OPTIMIZATION; PROBABILITY DENSITY FUNCTION; TELEPHONE SWITCHBOARDS;

DISCRIMINATIVE TRAINING; HIDDEN MARKOV MODELS; I-SMOOTHING; MAXIMUM MUTUAL INFORMATION; MINIMUM PHONE ERROR; MINIMUM WORD ERROR; OBJECTIVE FUNCTIONS; WORD ERROR RATE;

SPEECH RECOGNITION;

EID: 0036296863     PISSN: 15206149     EISSN: None     Source Type: Conference Proceeding    
DOI: 10.1109/icassp.2002.5743665     Document Type: Conference Paper

References (11)

1. Maximum mutual information estimation of hidden Markov model parameters for speech recognition
2. Decoder selection based on cross-entropies
3. An inequality for rational functions with applications to some statistical estimation problems
4. Prototype-based minimum classification error/generalised probabilistic descent training for various speech units
5. Hidden markov models, maximum mutual information estimation and the speech recognition problem
6. Minimum classification error rate methods for speech recognition
7. Improved discriminative training techniques for large vocabulary continuous speech recognition
8. Discriminative training for continuous speech recognition
9. MMIE training of large vocabulary speech recognition systems
10. Large scale discriminative training of hidden markov models for speech recognition