Refactoring acoustic models using variational expectation-maximization.
Pierre L. DogninJohn R. HersheyVaibhava GoelPeder A. OlsenPublished in: INTERSPEECH (2009)
Keyphrases
- expectation maximization
- acoustic models
- image segmentation
- em algorithm
- speech recognition
- hidden markov models
- mixture model
- automatic speech recognition
- maximum likelihood
- probabilistic model
- speech recognizer
- discriminative training
- broadcast news
- speaker independent
- generative model
- probability density function
- language model
- image processing
- bayesian framework
- maximum a posteriori
- gaussian mixture model
- pattern recognition
- information extraction