Adding Connectionist Temporal Summarization into Conformer to Improve Its Decoder Efficiency For Speech Recognition.
Nick J. C. WangZongfeng QuanShaojun WangJing XiaoPublished in: CoRR (2022)
Keyphrases
- speech recognition
- language model
- hidden markov models
- automatic speech recognition
- speech signal
- pattern recognition
- speech recognizer
- speech recognition technology
- speech processing
- speech understanding
- speech synthesis
- neural network
- noisy environments
- speech retrieval
- speech recognition systems
- speaker identification
- motion estimation
- speaker recognition
- keyword spotting
- speech recognition errors
- isolated word
- speech recognizers
- cepstral coefficients
- feature selection
- computer vision
- information retrieval