Joint Optimization of Spectro-Temporal Features and Deep Neural Nets for Robust Automatic Speech Recognition.
György KovácsLászló TóthPublished in: Acta Cybern. (2015)
Keyphrases
- neural nets
- joint optimization
- automatic speech recognition
- speech recognition
- noisy environments
- feed forward
- neural network
- artificial neural networks
- speech signal
- hidden markov models
- back propagation
- sparse representation
- spoken words
- conversational speech
- speech retrieval
- learning tasks
- non stationary
- multiscale
- word error rate
- genetic algorithm
- signal processing
- broadcast news
- spontaneous speech
- training data
- speech corpus