3M: Multi-loss, Multi-path and Multi-level Neural Networks for speech recognition.
Zhao YouShulin FengDan SuDong YuPublished in: CoRR (2022)
Keyphrases
- speech recognition
- multipath
- neural network
- pattern recognition
- end to end
- routing algorithm
- hidden markov models
- speech synthesis
- language model
- speech processing
- speaker independent
- automatic speech recognition
- speech recognition systems
- multi path routing
- noisy environments
- speech recognizer
- multiple paths
- speech recognition technology
- speaker identification
- speech signal
- wireless ad hoc networks
- multiresolution
- digital libraries
- feature extraction
- wavelet transform
- feature vectors
- speaker dependent