Gramazio Kohler Research

This research presents a design approach that uses machine learning to enhance architects’ design experience. Nowadays, architects and engineers use parametric design software (e.g. Grasshopper) to generate, simulate, and evaluate multiple design instances. In this project, we propose a Conditional Autoencoder that reverses the parametric modelling process and allows architects to define desired properties in their designs and obtain multiple predictions of designs that fulfil them. The results found by the encoder oftentimes goes beyond the user's expectations, thereby increasing the understanding of the design task and thus stimulating the design exploration.
Our tool also allows the architect to underdefine the desired properties to give additional flexibility in finding interesting solutions.
As a proof of concept, we used this tool within the architectural project Semiramis, a multi-story structure built in 2022 in the Tech Cluster Zug, Switzerland.

Source code: Visit the link.

Publications:

Luis Salamanca, Aleksandra Anna Apolinarska, Fernando Pérez-Cruz, Matthias Kohler, Augmented Intelligence for Architectural Design with Conditional Autoencoders: Semiramis Case Study. In: Design Modelling Symposium Berlin. DMS 2022: Towards Radical Regeneration pp. 108–121, 2022.
Link

Related projects:

AIXD: AI-eXtended Design

AI-Augmented Architectural Design

Credits:

Gramazio Kohler Research, ETH Zürich

In Zusammenarbeit mit: Dr. Luis Salamanca Mino, Prof. Dr. Fernando Perez Cruz (Swiss Data Science Center - SDSC)
Mitarbeiter: Dr. Aleksandra Anna Apolinarska (Projektleitung), Dr. Aleksandra Anna Apolinarska, Prof. Matthias Kohler


	Copyright 2024, Gramazio Kohler Research, ETH Zurich, Switzerland	[Press Download]


Gramazio Kohler Research Professur für Architektur und Digitale Fabrikation ETH Zürich HIB E 43 Stefano-Franscini Platz 1 / CH-8093 Zürich +41 44 633 49 06		Follow us on: Vimeo \| Instagram