Unlearning Intimacy

tags: future, gesture


Unlearning Intimacy is a meditation on the limits of machine learning. A StyleGAN machine learning network tries to understand the intangible emotional experience of intimacy vis a vis a physical embrace. Generative adversarial netwworks (GANs) typically rely on large self-similar datasets to create realistic results; this animation is generated through the processing of two deliberately small and unlike datasets: rendered motion capture data recording the physical act of hugging, and Google image search results from the search term “hug”.