Maximum-likelihood updates of HMM duration parameters for discriminative continuous speech recognition.
Rathinavelu ChengalvarayanPublished in: ICSLP (1998)
Keyphrases
- maximum likelihood
- continuous speech recognition
- discriminative training
- maximum likelihood estimation
- parameter estimation
- hidden markov models
- expectation maximization
- likelihood function
- em algorithm
- parameter learning
- hyperparameters
- maximum a posteriori
- maximum likelihood estimator
- gaussian mixture model
- genetic algorithm
- parameter estimates
- pairwise
- data sets
- posterior probability
- gaussian distribution
- closed form
- mixture model
- unsupervised learning
- object detection
- handwritten word recognition
- expectation maximisation
- log likelihood function