“Name That Move: Movement, Machine Learning, and Interactive Performance”

Abstract: How can bodily movement become material for interactive sound and visuals? In this talk, dance artist and data scientist Rose Xu 徐子瑜 will introduce Name That Move, an open-source toolkit for recording wearable motion-sensor data, training personalized movement classifiers, and connecting live predictions to creative systems such as TouchDesigner. Drawing from her research in dance, machine learning, and new-media performance, Rose will trace the project’s development from a concert-performance system into a reusable tool for artists and researchers.

The session will include a live demonstration and opportunities for audience interaction, while also exploring the creative decisions involved in teaching a machine a movement vocabulary. No programming experience is required. This work grew from Rose’s PhD research in the Department of Digital Arts and Experimental Media (DXARTS), and was developed into an open-source toolkit through two quarters of collaboration with Bryna Hazelton in the UW eScience Institute’s Data Science Accelerator program.

Speaker Bio: “Rose” Ziyu Xu (徐子瑜) is a researcher, choreographer, and performer working at the intersection of dance, new media arts, and machine learning. Rejecting the dichotomy of art and science, she explores how shared human questions can be articulated across different disciplinary languages. Rather than complying with AI to generate or replace creative labor, Rose designs artist-centered machine learning systems that deeply listen and respond to lived, embodied experience. Over the past years, Rose’s practice evolved from multimedia dance theater toward interactive systems integrating extended reality and computational intelligence, using the stage as a site to prototype future relationships between bodies, media, and computation. Across these works, Rose investigates how digital systems might engage embodied experience without erasing agency, ambiguity, or care.