Sequence Noise Injected Training for End-to-end Speech Recognition.
George SaonZoltán TüskeKartik AudhkhasiBrian KingsburyPublished in: ICASSP (2019)
Keyphrases
- end to end
- speech recognition
- noisy environments
- wall street journal corpus
- hidden markov models
- noisy speech
- acoustic models
- isolated word
- language model
- automatic speech recognition
- speech signal
- speech recognition technology
- pattern recognition
- speech synthesis
- noise level
- speech processing
- background noise
- congestion control
- speaker identification
- speech recognizer
- speech recognition systems
- noise reduction
- probabilistic model
- speaker dependent
- discriminative training
- speech recognizers
- noise model
- audio visual speech recognition
- training process