Log-Linear Optimization of Second-Order Polynomial Features with Subsequent Dimension Reduction for Speech Recognition.
Muhammad Ali TahirRalf SchlüterHermann NeyPublished in: INTERSPEECH (2011)
Keyphrases
- speech recognition
- dimension reduction
- feature extraction
- speech recognition systems
- log linear
- high dimensionality
- language model
- feature space
- hidden markov models
- automatic speech recognition
- pattern recognition
- high dimensional
- principal component analysis
- log linear models
- higher order
- linear discriminant analysis
- feature set
- unsupervised learning
- singular value decomposition
- low dimensional
- high dimensional data
- cluster analysis
- feature vectors
- neural network
- data sets
- nearest neighbor
- probabilistic model
- pairwise
- similarity measure
- machine learning