Hyung-Seok Oh

zero-shot learning (1) variational inference (1) self-supervised learning (1) speech synthesis (1) hierarchical model (1) self-supervised representation (1) cross-attention mechanism (1) text-to-speech synthesis (1) adversarial network (1) speech generation (1) emotional style control (1) spherical emotion vector (1) emotional intensity modeling (1) large language model (1) voice style transfer (1) filler insertion (1) filler prediction (1) filler style control (1) conversational speech generation (1)