Deep neural network features and semi-supervised training for low resource speech recognition.
Samuel ThomasMichael L. SeltzerKenneth ChurchHynek HermanskyPublished in: ICASSP (2013)
Keyphrases
- speech recognition
- supervised training
- neural network
- speech recognition systems
- hidden markov models
- pattern recognition
- speech recognizer
- speech signal
- automatic speech recognition
- language model
- speaker identification
- training algorithm
- speech synthesis
- feature set
- neural nets
- data mining
- artificial neural networks
- image processing
- machine learning
- structural features
- conditional random fields
- classification accuracy