Speeding up softmax computations in DNN-based large vocabulary speech recognition by senone weight vector selection.
Yingke ZhuBrian MakPublished in: ICASSP (2017)
Keyphrases
- speech recognition
- weight vector
- speech recognizer
- hidden markov models
- speech recognition systems
- language model
- pattern recognition
- learning rate
- speech signal
- hyperplane
- automatic speech recognition
- speaker adaptation
- linear combination
- speech recognition technology
- speech synthesis
- speaker independent
- speaker identification
- maximum likelihood
- support vector machine