Regularized feature-based maximum likelihood linear regression for speech recognition.
Mohamed Kamal OmarPublished in: INTERSPEECH (2007)
Keyphrases
- linear regression
- speech recognition
- maximum likelihood
- least squares
- speaker adaptation
- speech processing
- hidden markov models
- pattern recognition
- regression methods
- automatic speech recognition
- language model
- speech synthesis
- speech signal
- noisy environments
- speech recognizer
- ridge regression
- gaussian mixture model
- speech recognition systems
- nonlinear regression
- em algorithm
- expectation maximization
- speech recognition technology
- discriminative training
- hyperparameters
- speaker identification
- maximum a posteriori
- neural network
- gaussian distribution
- mixture model
- speaker diarization
- signal processing
- speaker independent
- feature vectors
- speaker dependent
- information retrieval