Thesis - Synergy

Group: Jing Zheng | Kurniati Kandiawan | Tyler Van Kirk
MATIAS DEL CAMPO | SANDRA MANNINGER
WINTER 2019

The relationship between artificial intelligence and architects has been a prevalent topic in architectural discourse. How large of a role can artificial intelligence play in the design of spaces? Is artificial intelligence simply a tool, or can it be a part of a collaborative design team composed of human and machine? These are questions designers need to consider before utilizing artificial intelligence in the design field. Currently, there is not enough advancement for artificial intelligence to have design sensibilities nor can this project attempt to claim the use of true artificial intelligence. Instead, this project intends to serve as a case study of how spaces can be co-designed by machine learning algorithms and designers. What contradictions may occur between what is intended and what the algorithm produces? Do these conflicts need to be resolved or do they prevent a better alternative that challenges the designers to own sensibilities?

In the process of this project, the proximity and limitations are inputted into the machine learning algorithm according to the desired design allowing the algorithm to generate a form. If the output is not as desired, the initial inputs can be modified and the process can be run again. Through this recursive process, a better understanding of the nature of the algorithm is gained allowing one to be more informed on how to manipulate the inputs to result in a similar if not better result than the original intent.

Although this project cannot claim to perfect the relationship between artificial intelligence and designers, it contributes to the ongoing conversation of what the potential role of artificial intelligence in design can be and attempts to discover new techniques and work methods to better utilize AI for the future of the design profession

Presentation Slides of Synergy

These images are generated through Deep Dream & Style Transfer using Google Deep Dream Generator. It is an exploration on Machine Learning engines available online, showing the results of the learning process of the Artificial Intelligence and it's application to the inputted texture and design. These texture will then be further explored and applied in Grasshopper and Rendering in the later process of the thesis. 

Final design and rendering selection from the exploration.

We were also exploring the possibility of the same Machine Learning in a micro-level design application, such as furniture. This chair is one of the result of the machine learning, which is then fabricated for the final presentation of the thesis. 

Autodesk Maya works.

Here, we explore the usage of a different 3D design software, reimagining what the multiplication and mirroring effect of a simple 3D object can create to form a tower. Starting to explore making a façade and slowly, these form spreads to create a complex and intricated spaces within as well.