Ambiguous forms are created from the merging and morphing of multiple heterogeneous inputs. The Machine Learning model, Stylegan2, has the capacity to take collections of disparate images and find some common form within the dataset. The resulting images are not entirely a mixture or merging of them various inputs (A+B≠AB), but something wholly new (A+B=C).
This project leverages the generative capabilities of Stylegan2 to explore a new aesthetics that emerges when multiple datasets collide. Datasets of Brutalist architecture, Yugoslavian war monuments, and Greek statues are combined to train a Stylegan2 model. Each of these was chosen for their use of material and for the formal language they use. The resulting imagery does not quitem resemble any of the three inputs. It is a series of ambiguous forms.
