Grapheme-to-phoneme conversion based on adaptive regularization of weight vectors.
Keigo KuboSakriani SaktiGraham NeubigTomoki TodaSatoshi NakamuraPublished in: INTERSPEECH (2013)
Keyphrases
- adaptive regularization
- weight vector
- grapheme to phoneme conversion
- parameter selection
- natural language processing
- learning rate
- super resolution
- image deblurring
- language resources
- total variation
- linear combination
- hyperplane
- objective function
- regularization parameter
- blind image deconvolution
- regularization term
- support vector machine
- loss function
- feature values
- principal components
- model selection
- data sets
- convergence rate
- image restoration
- natural images
- training data
- computer vision