How does Picbreeder work?
Picbreeder generates a population of images that a user can choose. The selected images are then used to evolve a new generation via Compositional Pattern Producing Networks (CPPNs). CPPNs are neural networks that contain activation functions which allow them to create and evolve images.
Who created it?
The original idea for Picbreeder came from Jimmy Secretan, who worked with Nick Beato, David D'Ambrosio, Adam Campbell, Adelein Rodriguez, and Dr. Kenneth Stanley to implement it. In 2020, the Picbreeder website was revitalized by Cameron Henkel, Nicholai Benko, Jaden Williams, Sarah Friday, Anna Krolikowski, and Bennet Leff.