Diffusion (2006) is a series of digital sketches that presents an unconventional approach to portraying communities as well as an innovative approach to digital mark-making. The drawing gains its unique figure from implementing a kind of neural network. A neural network is a machine-learning method that simulates a system of communicating nerve cells. In Diffusion, the simulated community responds to input via mouse position. Each pixel has a limited memory of what brightness it has been and of the brightness of the pixels around it. The system is constantly in a state of averaging and diffusing any changes that the viewer creates on the screen. To the user, the experience is much like dripping colored water into a puddle. The elegant effect is to diagram how data spreads through an interconnected community of modular cells.