C.O.P.S # 1, 2 (Mold / AI graphic)

art work title:
C.O.P.S.
Collection of primitive steps
versions 1, 2

date of creation:
apr 2020

expo:
Mozilla Hub
“Expectation of the new nature”, 2020;

description:
The biological system develops by making primitive decisions. The main tasks of such system are nutrition and reproduction. During implementation, a system makes simple, but at the same time effective solutions that ensure the fastest distribution in search for nutrients and the reproduction process.

A colony of zygomycete was taken as an example of a bio-system. The colony’s development is polycentric, which forms the thesis about the distribution of the decision-making process in the biological system. The project explores the behaviour pattern of a mushroom colony through algorithmic analysis.

description:
The project continues the research on biological systems using algorithmic analysis methods. The variety of fungal colonies becomes a dataset for a neural network.

In the first case, a pre-trained model of neural network is used, in the second, an own model is created and trained.
Total neural network training time: 20 hours.
Number of training stages (snapshot): 38.

A dataset, which was created for the” Collection ” from mold images, was automatically found with the “Bing” search engine. The StyleGAN 2 neural network processed more than 300 images, creating new types of mold colonies.

Project vers. 1 was produced on “Materials” intensive at “–Da.Digital art” 
Master Program FEFU with tutor Maria Kuptsova.
Project vers. 2 was produced on “Neural Networks” intensive at “–Da.Digital art” 
Master Program FEFUwith tutor Artem Konevskikh.

Last update 08 May 2021