Speaker adaptation of neural network acoustic models using i-vectors.
George SaonHagen SoltauDavid NahamooMichael PichenyPublished in: ASRU (2013)
Keyphrases
- speech recognition
- acoustic models
- speaker adaptation
- speaker independent
- neural network
- speech recognizer
- automatic speech recognition
- hidden markov models
- pattern recognition
- speaker identification
- spoken language
- digit recognition
- speech signal
- broadcast news
- feature vectors
- maximum likelihood
- language model
- speech recognition systems
- training process
- feature selection