An empirical study of learning rates in deep neural networks for speech recognition.
Andrew W. SeniorGeorg HeigoldMarc'Aurelio RanzatoKe YangPublished in: ICASSP (2013)
Keyphrases
- speech recognition
- learning rate
- neural network
- pattern recognition
- backpropagation algorithm
- learning algorithm
- convergence rate
- automatic speech recognition
- language model
- speech processing
- speech synthesis
- hidden markov models
- speech signal
- speech recognizer
- back propagation
- noisy environments
- speech recognition technology
- covering numbers
- speech recognition systems
- genetic algorithm
- uniform convergence
- speaker identification
- convergence speed
- artificial neural networks
- gaussian kernels
- sample complexity
- training process
- speaker independent