A sunless summer in Shangri La
A sunless summer in Shangri La, 2022
Video and Artificial Intelligence.
Matheus da Rocha Montanari
In "A sunless summer in Shangri Lá", the weather becomes one of the artistic agents, confronting the utopian literary city high up in Tibet, the Brazilian beach on the coast of Rio Grande do Sul, and the mountains of the Yosemite Park in the United States.
Shangri Lá was originally created in 1933 by James Hilton in his novel Lost Horizon. It is an imaginary and mountainous place, supposedly in the region of Tibet, where the inhabitants never aged, as long as they never left, oscillating between a paradise and a prison.
The work operates by tensioning relationship between the natural and the artificial. Among grains of sand and grains of pixelated noise, cyclic adversarial generative networks (cycle GAN) are used to construct an imagistic set.
The artificial intelligence networks used in the project are trained with images from different seasons in the same location. From this data set, they compete with each other to artificially alter the images with opposing seasons, until one convinces the other of the naturalness of its artificialization.
The model was developed based on images from the Yosemite Park in the United States of America. The park is often visited by professional and amateur climbers who venture out on climbs, registering the routes in reports, photos, videos, and blog posts. A well-known climber from the region, Bruce Bindner, or Brutus of Wyde, as he was known, claims to have found Shangri Lá in one of his expeditions. Brutus described, and even made some drawings of the beautiful place in the High Sierra, but kept its location a secret.
The algorithm is trained with a set of 853 images of the park in the summer and 1273 images in the winter, as reference. From these, it translates any input image to the opposite season of the year. By doing so, the definitions of winter and summer, despite their globality and diversity, are established based on the local characteristics of the park.
We confront this idea by introducing images of the Xangri Lá beach in a cloudy summer day. Without references for dunes of sand, the algorithm transforms them into mountains covered in snow. This creates the utopian landscape of a sub-tropical snowy beach. Unknowingly, the system replicates the seasons of the northern hemisphere in the southern hemisphere, creating dreamlike scenes in search of the sun of Shangri Lá.