CASS-NAT: CTC Alignment-Based Single Step Non-Autoregressive Transformer for Speech Recognition.
Ruchao FanWei ChuPeng ChangJing XiaoPublished in: ICASSP (2021)
Keyphrases
- speech recognition
- autoregressive
- single step
- non stationary
- multi step
- hidden markov models
- spectral analysis
- speech recognition technology
- speech signal
- language model
- pattern recognition
- speech recognizer
- automatic speech recognition
- random fields
- speaker identification
- speech synthesis
- sar images
- linear prediction
- speech recognition systems
- noisy images
- signal processing
- probabilistic model