2.0 Fluid Body

presents an intercultural co-adaptation, where ballet postures do not imitate Guzheng technique, nor does the instrument override the logic of dance.

YUTING XUE
YUESHEN WU
ELKE REINHUBER

This work proposes an intercultural installation mediated by supervised machine learning, in which the embodied vocabulary of classical ballet serves as a generative interface for performing the aesthetics of Guzheng. Through a combined classification–regression model coupled with a granular synthesis engine, the system learns how the bodily intention, spatial orientation, and dynamic qualities of ballet posture shape sonic expression, enabling performers to play sound textures of the Guzheng in real time through movement. Within this feedback loop, movement and sound continually reshape each other, forming a dynamic field in which distinct Eastern and Western cultural traditions engage through the body.

In this way, Fluid Body presents an intercultural co-adaptation, where ballet postures do not imitate Guzheng technique, nor does the instrument override the logic of dance. Instead, their interaction, mediated by machine learning, creates a space in which two cultural forms remain responsive and relational. This practice- based research demonstrates how creative machine learning mediation can cultivate new forms of artistic translation, revealing the potential of embodied practice to generate shared aesthetic meaning across cultural boundaries.