Albeit this might seem hoary and formuliac, the artwork I have submitted focuses on the seemingly haphazard nature of particles (quantum mechanics) and how when examined at the appropriate distance, they are quite geometrically and symmetrically congruous (e.g. the theory of relativity.)
Imagine you have a big circle (circumference, not disc) around the origin, which consists of a "sticky" border. You start a simple two-dimensional random walk from the origin; the particle moves until it touches one of the sticky particles, and then it stops, and itself becomes "sticky".
Then you start the second particle from the origin, which does exactly the same thing, until it get stuck, and so on. Eventually, these sticky particles start forming "stalactites" and the process stops when the origin (0,0) itself becomes a sticky particle and we cannot continue.
What you see in the picture is the final stage of this process.
Chaotic Systems: Odesa is a generative multimedia installation based on chaotic dynamical systems and their sensitivity to initial conditions (commonly referred to as the "butterfly effect"). At its core are two real-time Chua attractor generators, built with STM32 microcontrollers, solving differential equations with nearly identical starting parameters—except for one infinitesimal variation (picorange). This difference grows over time, making the systems diverge unpredictably.
The divergence values are sent to a conditional generative adversarial network (cGAN), trained on two datasets: peaceful photographs of Odesa and archival imagery from violent street clashes in 2014. As divergence increases, the neural network generates increasingly chaotic imagery. Once a specific threshold is reached, the system resets, beginning a new cycle.
Through autonomous computation and generative media, the work draws connections between abstract mathematical models and real-world instability. It suggests that imperceptible variations—whether in physical, biological, or social systems—may trigger dramatic transformations. The work runs continuously, using two embedded computers (Orange Pi) and three displays for video, attractor plots, and system graphs.
By integrating mathematics, AI, and historical context, the project highlights hidden mechanisms that shape the visible world.