What’s On You and AI: Through the Algorithmic Lens

Counting Craters on the Moon

Kyriaki Goni

Courtesy of Mohamad Ali-Dib and the DeepMoon team

Counting Craters on the Moon, 2019 Kyriaki Goni

The video, part of the homonymous multimedia installation, presents an imaginary encounter between acclaimed Johann Friedrich Julius Schmidt (1825-1884) and neural network DeepMoon. DeepMoon was developed in March 2018 at the Canadian Institute for Theoretical Astrophysics University of Toronto. Julius Schmidt became in 1958, director of the new National Observatory of Athens in Greece, where the clear skies were very suited to astronomical observation. Both of them set out to count lunar craters, which due to the lack of atmosphere on the Moon remain unchanged for billions of years and thus provide valuable information for the past and possibly the future of our near solar system.

Julius Schmidt dedicated 35 years of his life and produced one of the most detailed lunar maps with around 30.000 craters, while DeepMoon spotted around 7.000 craters in less than 24hrs. In their dialogue they discuss precision, error, dedication, technique, nomenclature, loneliness and patterns.

The artist collaborated with the DeepMoon team, who kindly provided the dataset to be part of the installation and ran the neural network on Schmidt’s map, which the artist digitally stitched together from 25 separated sections.

Speculating upon the possible synergies between human and machine, this work is an invitation to imagine how we can learn from and with machines in order to build different, multiple and, possibly, collective understandings of the surrounding world and its cosmos.

Title: Counting Craters on The Moon

Medium: multimedia installation

Artist: Kyriaki Goni

Year: 2019

Location: On display at Pedion tou Areos

Glossary: neural network