Phoneme Segmentation Using Self-Supervised Speech Models

Luke Strgar, David Harwath

We apply transfer learning to the task of phoneme segmentation and demonstrate the utility of representations learned in self-supervised pre-training for the task. Our model extend…