In this post, I present the results of a first test on the application of a class of AI techniques – named Generative Models – to construct design spaces from a dataset of design precedents. The aim is showing a new approach in computational design that moves beyond problem-solving. Here AI is proposed as a tool that can support design thinking and creativity.

Artificial Neural Networks have long been applied in structural design for real-time structural analysis and optimisation. Although such applications have demonstrated the possibility to achieve high computational efficiency with a negligible loss in accuracy, they have only scratched the surface of the true potential of these models in computational design. As part of my PhD, I’m investigating how Artificial Intelligence can be leveraged during the development of design ideas, that is when considerations about design requirements are more relaxed. In this post, I present the results of a first test on the application of a class of AI techniques – named Generative Models – to construct design spaces from a dataset of design precedents. The aim is showing a new approach in computational design that moves beyond problem-solving. Here AI is proposed as a tool that can support design thinking and creativity.

These results will be presented at the 1st Italian Workshop on Shell and Spatial Structures (IWSS 2020), which will take place on 25^th and 26^th June 2020 and will be broadcast online. Here you find a preview of the presentation that I will give during the workshop. Enjoy!