Data from every scale of design development – from figure ground to façade and material distribution – is used within the design process.

The generative adversarial network (GAN) is an unsupervised example-based machine learning algorithm that can ‘quantify’ the style of iconic buildings and transfer these quantifications to other iconic buildings.

The programmes of recycling and cultural facilities are blended together through the use of the GAN algorithm.

A dataset of images of Brutalist architecture are used to train a machine learning algorithm to generate a building’s façade.

A view illustrating the blending of pixel data from Brutalist cultural architecture together with recycling façade patterns.

The building elevations and façades adapt the same design logic. The data are extracted from the GAN results. The blocks on the façade are made of various recyclable materials, which form a view appearing to ‘glitch’ and reveal the complexities in architecture.