Prototype-based minimum classification error/generalized probabilistic descent training for various speech units.
Erik McDermottShigeru KatagiriPublished in: Comput. Speech Lang. (1994)
Keyphrases
- minimum classification error
- discriminative training
- statistical pattern recognition
- probabilistic interpretation
- hidden markov models
- maximum likelihood estimation
- classification error
- posterior probability
- gaussian mixture
- speech recognition
- automatic speech recognition
- probabilistic model
- pattern recognition
- data sets
- speech signal
- bayesian networks
- generative model
- maximum likelihood
- probabilistic inference
- multi view
- decision trees