Fast model selection based speaker adaptation for nonnative speech.
Xiaodong HeYunxin ZhaoPublished in: IEEE Trans. Speech Audio Process. (2003)
Keyphrases
- model selection
- speaker adaptation
- speech recognition
- automatic speech recognition
- maximum likelihood
- cross validation
- speech recognizer
- speaker dependent
- hyperparameters
- speaker independent
- speech signal
- regression model
- parameter estimation
- vocal tract
- sample size
- speech synthesis
- machine learning
- variable selection
- model selection criteria
- selection criterion
- speaker identification
- hidden markov models
- mixture model
- gaussian process
- parameter determination
- pattern recognition
- statistical inference
- spoken language
- unsupervised learning
- automatic model selection
- data mining
- feature selection
- information criterion
- bayesian information criterion
- broadcast news
- noisy environments
- marginal likelihood
- speaker recognition
- language model