Artists
Max Brodeur
Max Planck Institute for Brain Research
Frankurt, Germany
Statement
Simple dynamical systems often give rise to beautifully complex structures—a great opportunity for artistic visualizations. I’m particularly interested in brains, artificial or biological, and the elegant mathematics that govern them.
Artworks
The brain is a high-dimensional dynamical system that sometimes operates at the edge of chaos—between stability and instability. The precise role this critical state plays in neural computation remains a mystery.
In this work, the edge of chaos reveals diverse structures emerging from the dynamics in a random recurrent network with 1000 neurons and connectivity weights sampled from a Gaussian distribution.
In a 5x5 grid, each row corresponds to a distinct network initialization. The activity is projected to lower dimensions with Principal Component Analysis (PCA), and in the columns are the first ten principal component dimensions of the activity, plotted against each other in consecutive pairs (e.g. PC1 vs. PC2, to PC9 vs. PC10).