Exploiting semi-supervised training through a dropout regularization in end-to-end speech recognition.
Subhadeep DeyPetr MotlícekTrung BuiFranck DernoncourtPublished in: CoRR (2019)
Keyphrases
- end to end
- speech recognition
- supervised training
- hidden markov models
- conditional random fields
- training phase
- language model
- unsupervised learning
- supervised learning
- pattern recognition
- dependency parsing
- neural nets
- training algorithm
- speech recognition systems
- speech signal
- recognition process
- speech recognizer
- e learning
- expectation maximization
- principal component analysis
- feature space
- training data
- information retrieval
- neural network