Utterance Invariant Training for Hybrid Two-Pass End-to-End Speech Recognition.
Dhananjaya GowdaAnkur KumarKwangyoun KimHejung YangAbhinav GargSachin SinghJiyeon KimMehul KumarSichen JinShatrughan SinghChanwoo KimPublished in: INTERSPEECH (2020)
Keyphrases
- speech recognition
- end to end
- wall street journal corpus
- isolated word
- hidden markov models
- language model
- pattern recognition
- speech recognition technology
- speech recognizer
- acoustic models
- speech signal
- speech synthesis
- speech processing
- congestion control
- automatic speech recognition
- speech recognition systems
- noisy environments
- feature selection
- training process
- signal processing
- speaker dependent
- probabilistic model
- speech recognizers
- feature space
- face recognition