Material Intelligence in the Circular Economy: The End of Waste

Carlos Navarro, Jiaoyue Zhao

M. DesR SCI-Arc, 2015, and M. Arch II SCI-Arc, 2020

The workshop explores the agency of neural networks in the study of material stages within the context of life-cycle assessment in the circular economy. It primarily focuses on the use of GAN (generative adversarial networks), in Runway ML and Google Colab, to identify new cradle-to-cradle opportunities within latent walks from material life-cycle image datasets gathered from conventional linear economy industries’ take-make-waste processes. This workshop engages in the instrumentality of AI in recycling and closing life-cycle loops, not only in terms of proper post-Carbon waste management but also in potential manufacturing approaches. Generatively, the workshop seeks to prompt the use of 3D GAN techniques into animation and simulation industry-standard software – Cinema 4D –, through the translation of 2D GAN outputs as three-dimensionalized dynamic models to extract actual volumetric features of newly informed material stages, providing further opportunities for digital physicalization. Visualization-wise, the workshop covers methodologies to construct subtly animated novel ecologies with Cinema 4D and Quixel Bridge, as a platform for a participatory generative design that combines human-curated and crafted photo-realistic assets and datasets with AI volumetric processing of architectural artifacts. The resulting outputs are expected to lie within AI indeterminacy and human control while challenging and augmenting the designer’s agency. The workshop grounds itself on recent widespread knowledge in the field of artificial intelligence applied to architectural design. This workshop offers a brand-new approach in the generation of 3D GAN by utilizing robust procedural 3D animation software to produce fully volumetric models, while still grounded in the processing of accessible 2D GAN outputs. This newly animated AI-based approach extrapolates the opportunities for the digital physicalization of generative datasets, granting them a material fluidity of identifiable stages prone to be problematized within the context of life-cycle assessment in the circular economy and SDG12 (on Sustainable Consumption and Production). GANs in RunwayML and Google Colab for finding cradle-to-cradle checkpoints within waste-based latent walks and machine vision Processing scripts for automated waste imagery recycling distribution are key tools that the workshop provides for the generation of personalized input datasets and corresponding digitally recycled outputs.