Discriminative training for speech recognition is compensating for statistical dependence in the HMM framework.
Dan GillickSteven WegmannLarry GillickPublished in: ICASSP (2012)
Keyphrases
- speech recognition
- hidden markov models
- discriminative training
- automatic speech recognition
- speech signal
- language model
- statistical dependence
- pattern recognition
- acoustic models
- speech recognizer
- speech recognition systems
- neural network
- learning algorithm
- probabilistic model
- speaker identification
- speaker adaptation