Prototype-based minimum error training for speech recognition.
Erik McDermottShigeru KatagiriPublished in: Appl. Intell. (1994)
Keyphrases
- speech recognition
- minimum error
- wall street journal corpus
- isolated word
- hidden markov models
- language model
- acoustic models
- pattern recognition
- speech signal
- speech processing
- automatic speech recognition
- speech recognition technology
- speech synthesis
- speech recognition systems
- training process
- speech recognizer
- bayes optimal
- noisy environments
- speaker identification
- classification error
- training set
- speaker dependent
- supervised learning
- speaker independent
- training examples
- speaker adaptation
- feature set
- information retrieval