"The Automated Landscape" entered a new stage of automation.

Walking through a Japanese city, one encounters familiar designs and repeating interfaces. I wanted to see if I can build a neural network that can create new "automated landscapes". 

I created a machine learning model and let it produce new images of this kind of interface. A machine learning model is a file that has been trained to recognize certain types of patterns. You train a model over a set of data, providing it an algorithm that it can use to reason over and learn from those data.

The video above was created with a dataset of 400 images. I trained StyleGAN2-ADA in Google Colab for roughly three days.

Go to the Instagram Account of The Automated Landscape to see more.