, Data Visualization Researcher, Northeastern University
Artists are uniquely positioned to lend new approaches to curating training data and post-processing treatment of outputs for novel creations with models such as StyleGAN. Through the use of generative design, batch editing, unusual application of machine learning transformations, and transmedia augmentations, artists can "think outside the box" when preparing training data. The potential avenues for daisy-chaining models in a creative workflow are combinatorially innumerable, and worthy of documentation and study. Additionally, there's a fascinating emergent landscape of media study regarding integrating model outputs into broader artistic processes, such as physicalization, interactivity, transmedia storytelling, nonlinear narrative capability, and subjective expression of human experience.