Prior knowledge guided maximum expected likelihood based model selection and adaptation for nonnative speech recognition.
Xiaodong HeYunxin ZhaoPublished in: Comput. Speech Lang. (2007)
Keyphrases
- model selection
- speech recognition
- prior knowledge
- speaker adaptation
- likelihood function
- selection criterion
- cross validation
- hidden markov models
- hyperparameters
- language model
- automatic speech recognition
- speech synthesis
- maximum likelihood
- pattern recognition
- speech signal
- information criterion
- mixture model
- speech recognizer
- model selection criteria
- machine learning
- noisy environments
- variable selection
- feature selection
- gaussian process
- speaker identification
- unsupervised learning
- automatic model selection
- speech recognition technology
- bayesian information criterion
- marginal likelihood
- speech recognition systems
- training data
- semi supervised
- high dimensional
- speaker independent
- decision trees
- speaker dependent