Sarah Taylor

Papers

Total Citations

H-Index

About

Dr. Sarah Taylor is a leading researcher in embodied conversational AI, whose work bridges the critical gap between speech and naturalistic nonverbal behavior. Her primary contributions lie in developing generative models for speech-driven animation and human-computer interaction, with a focus on creating agents that can both express their own discourse and react dynamically to interlocutors. Her influential 2021 paper, "Speech-Driven Conversational Agents using Conditional Flow-VAEs," introduced a novel Flow Variational Autoencoder framework that learns the complex, multimodal mapping between speech acoustics and full-body gesture motion. This work, which has garnered 17 citations, addresses the fundamental challenge of enabling virtual agents to move expressively while responding naturally to incoming speech in real-time. Dr. Taylor’s research has direct applications spanning animation, robotics, and interactive systems, and her methodological innovations in conditional flow-based models represent a significant advance in the field of embodied communication. Her work continues to shape how researchers approach the automatic generation of coordinated speech and gesture in autonomous agents.