The form boundaries do not exist in the latent space learned by an artificial neural network. The assemblage of parts is reconstructed in new ways each time a concept is generated. The proportions are skewed to fit a path between every two points. The artificially imagined classical order lives in the generalization of the principal components.
The artist is training AI (GAN) to find transformations and intersections of different discrete concepts, which is unnatural to humans as we imagine them in isolation. In their technique, a training collection of thousands of photographs of memories captured on film and matrix serves as guidance material to extract new meanings. The artist explores how the individuality of the original images is lost when they are imagined by GAN.
Most of Tau’s inspiration comes from experimental and surrealist photography. The shift to generate one’s own reality with a camera, instead of mimicking the surrounding world, allows dwelling on the numerous dimensions of the data surrounding us. In a very general sense, machine learning and artificial intelligence are just a new way of seeing things, from a very far perspective, not bound by the limits of our minds, used to low-dimensional information.
