Uconv-Conformer: High Reduction of Input Sequence Length for End-to-End Speech Recognition.
Andrei AndrusenkoRauf NasretdinovAleksei RomanenkoPublished in: CoRR (2022)
Keyphrases
- end to end
- speech recognition
- hidden markov models
- automatic speech recognition
- language model
- speech synthesis
- pattern recognition
- speech recognizer
- speech signal
- speech recognition systems
- speech processing
- speech recognizers
- congestion control
- speech recognition technology
- noisy environments
- speaker identification
- isolated word
- speaker adaptation
- speaker dependent
- neural network
- multi channel
- bayesian networks
- feature extraction
- image processing
- feature selection
- machine learning