wav2vec: Unsupervised Pre-training for Speech Recognition.
Steffen SchneiderAlexei BaevskiRonan CollobertMichael AuliPublished in: CoRR (2019)
Keyphrases
- speech recognition
- wall street journal corpus
- isolated word
- hidden markov models
- supervised learning
- acoustic models
- automatic speech recognition
- language model
- speech synthesis
- pattern recognition
- speech processing
- speech understanding
- speech signal
- speaker identification
- speech recognition systems
- speech recognizer
- noisy environments
- unsupervised learning
- semi supervised
- training set
- speech recognition technology
- keyword spotting
- speech recognition errors
- machine learning
- cepstral coefficients
- speaker recognition
- training process
- speech retrieval
- training data
- computer vision