Conformer-Based on-Device Streaming Speech Recognition with KD Compression and Two-Pass Architecture.
Jinhwan ParkSichen JinJunmo ParkSungsoo KimDhairya SandhyanaChangheon LeeMyoungji HanJungin LeeSeokyeong JungChangwoo HanChanwoo KimPublished in: SLT (2022)
Keyphrases
- speech recognition
- hidden markov models
- language model
- speech processing
- automatic speech recognition
- speech signal
- speech recognizer
- speech recognition technology
- pattern recognition
- speech synthesis
- noisy environments
- keyword spotting
- isolated word
- speech understanding
- compression algorithm
- speaker dependent
- speech recognition systems
- speaker identification
- speech retrieval
- feature space
- information retrieval
- compression ratio
- speaker independent
- image compression
- speech recognizers
- computer vision