life after digital

Neural networks were trained on different outputs by the artist (b/w drawings, polaroids, digital paintings) as well as stock photographs for an outside influence. A single artist's body of work without any evident or pertinacious style yields vastly different results than neural networks exclusively trained on a single phase of Vincent van Gogh or Claude Monet, and the networks struggle finding a clear path both when translating from trained and untrained images. This set is a floor installation of 225 images collected from different stages of the training process of three individual models. Many things apparently deemed unrecognizable seem to be wrapped in plastic; but a few night skies and sunsets feel appropriately placed. Somewhat correct skin tones appear and disappear again during the training process. While trees turn green, some lines inexplicably start glowing red. And while the models can produce new images like these indefinitely, as of yet, none of them has turned out to be a skilled curator. The artist went through the thousands of images created in this process and sorted them according to color, patterns, moods, and emerging themes, making it seem as if the models, at one point or another, for just a moment, had an idea, a plan, or a spark of creativity. They, of course, didn't.