The lecture will discuss the generation of three-dimensional architectural forms as one that is increasingly (and inevitably) about the sampling of forms in n-dimensions. Whether it is the implicit sampling of forms from the continuous latent space of generative adversarial networks (GANs) or the explicit sampling of forms from the discrete probabilistic sequence of autoregressive models; today, to design with generative AI models entails a computationally plastic conception of programming and producing forms. It is by no coincidence that the concept of plasticity could be found in modern art (Neoplasticism) as well as in today’s neuroscience (Neuroplasticity). By means of a series of past and recent projects using machine learning and deep learning, the theory of architectural sampling will be brought forth during the lecture.
Click HERE to register.