Matt Shannon

Latest

• Robust and Unbounded Length Generalization in Autoregressive Transformer-Based Text-to-Speech
• Learning the joint distribution of two sequences using little or no paired data
• Speaker Generation
• Non-Saturating GAN Training as Divergence Minimization
• Location-Relative Attention Mechanisms For Robust Long-Form Speech Synthesis
• Semi-Supervised Generative Modeling for Controllable Speech Synthesis
• Effective Use of Variational Embedding Capacity in Expressive End-to-End Speech Synthesis