Latent Displacement researches dimensional expressions navigating the latent space of a Compositional Pattern-Producing Network ( CPPN ). The generative piece explores the hidden beauty produced by the multi-dimensionality of a Recursive Neuronal Network.

On its default state, a new CPPN model is generated every ~5 seconds. By toggling randomization on the settings panel, the CPPN remains constant while altering two of its latent vectors sequentially.

The model's output produces an albedo and displacement map used to generate the base geometry.

Technology : WebGl, RNNs