On training the recurrent neural network encoder-decoder for large vocabulary end-to-end speech recognition.
Liang LuXingxing ZhangSteve RenalsPublished in: ICASSP (2016)
Keyphrases
- speech recognition
- end to end
- recurrent neural networks
- rate allocation
- recurrent networks
- speech recognizer
- speech recognition systems
- video codec
- hidden markov models
- low complexity
- distributed video coding
- neural network
- isolated word
- pattern recognition
- feed forward
- speech signal
- acoustic models
- language model
- speaker independent
- source coding
- error control
- speaker adaptation
- automatic speech recognition
- artificial neural networks
- training process
- turbo codes
- scalable video
- bit rate
- speaker identification
- wyner ziv
- discriminative training
- video coding
- rate distortion
- video compression