Dancing With Myself

Explore the intersection of dance, machine learning, and personal identity in this 36-minute conference talk from Strange Loop 2021. Discover how Mariel Pettee, a researcher at Lawrence Berkeley National Lab, used motion capture data of her own movements to train custom machine learning models, including a Variational Autoencoder (VAE) and Graph Neural Network (GNN), to generate choreography that mimics her unique style. Delve into the creative process, technical challenges, and philosophical questions that arose while developing AI as a dance partner during a year of physical distancing. Gain insights into the potential and limitations of algorithms in capturing and reflecting human identity through movement, and explore the broader implications of AI as a creative collaborator in the arts.