Effects of Architecture Choices on Sparse Coding in Speech Recognition.
Fionntán O'DonnellFabian TriefenbachJean-Pierre MartensBenjamin SchrauwenPublished in: ICANN (1) (2012)
Keyphrases
- speech recognition
- sparse coding
- dictionary learning
- sparse representation
- natural images
- image classification
- linear combination
- hidden markov models
- speech synthesis
- language model
- image representation
- speech recognizer
- pattern recognition
- generative model
- unsupervised learning
- automatic speech recognition
- speech recognition systems
- image patches
- signal processing
- input image
- visual words
- high dimensional
- speech signal
- bayesian networks
- speaker identification
- computer based instruction
- visual features
- machine learning
- small number
- feature space
- learning algorithm