with elina zazulya
Foraminiferal is an ephemeral, intermediate state between humans and the microorganism foraminifera, a microscopic inhabitant of predominantly salty waters, which has a calcareous porous body. The special surface structure of this creature provides it with a sensitive connection with the outside world and allows it to breathe throughout its body. Exploring the capabilities of neural networks, the artists connect humans with other forms of life and reflect on how the human body would transform if it moved from one habitat to another.

The visual part is a three-channel video of interpolations generated by the StyleGAN 2 neural network. The first stage of training the neural network took place on a set of images of the human body. Next, the resulting model was retrained on a set of images of microorganisms. To generate the final videos, intermediate stages were chosen, where something between the initial model (trained on the human body) and the formifera was generated. To generate sound, the RAVE neural network was used, which was trained on the sounds of deep, calm breathing of people, and then used to repeat the sound of waves with their breath.

The project was created with the support of the Art & Science Center of ITMO University, Aurora Scientific Communication and the Institute for Stratigraphic Research (Instituto de Investigaciones en Estratigrafía, IIES) of the University of Caldas (Universidad de Caldas), Manizales, Colombia.