Feature-based pronunciation modeling with trainable asynchrony probabilities.

Karen Livescu, James R. Glass

Browse the full Interspeech paper archive.