Computational Explorations 2

Deep generative models and reinforcement learning transform the architectural design landscape by creating diverse design options that respond to real-world constraints. This course invites students to explore how these advanced AI techniques can be harnessed in architectural design to generate creative possibilities and improve design outcomes.

Students will gain hands-on experience training, deploying generative models that create novel architectural forms and design elements, and designing reinforcement learning frameworks that iterate on architectural solutions.

Beyond technical skills, students will discuss the implications of generative AI for the future of architecture, exploring its potential and ethical considerations. Critical reflection on the limitations of these methods will be an integral part of the course, as students will examine issues such as the potential for bias in AI-generated designs, the challenges of ensuring design feasibility, and the ethical implications of replacing human creativity with automated processes.

The module includes lectures, tutorials, and projects where students apply their knowledge to design challenges. Assessment is based on weekly assignments and a final group project showcasing the application of generative models and/or reinforcement learning to a design problem.

The module assumes familiarity with deep learning concepts and Python programming, as well as experience with tools such as Rhino/Grasshopper. Prior successful participation in Computational Explorations 1 is mandatory.

Successful participation in the Computing in Architecture Seminar is mandatory for attending this seminar.