Tower GAN


Machine Intelligence in Architecture Workshop
Summer 2021
DigitalFUTURES

w/ Zhifei Xu







︎︎︎




“Tower GAN” is a project that studies tower as a typology. Towers, most of the time designed in pursuit of the iconic, are just like vases in the city. The design of a tower is usually shape-oriented, which makes it a good opportunity to study through AI, especially for 3d GAN.

Made using machine learning workflows.




Type:               Workshop
Institution:    DigitalFUTURES
Instructors:   Runjia Tian, George Guida, &
                         Dongyun Kim.
Research:      Machine Learning, 2D & 3D GAN.



1. StyleGAN model trained on all the different kinds of grid facades throughout history.
2. Real-time latent walk texture mapping.
3. 3D GAN model running.
4. Real-time 3d GAN model.
5. Project board.















CARLOS NAVARRO      LOS ANGELES, CA