Unplugging evolutionary algorithms: On the sources of novelty and creativity
Francisco Fernández de Vega, Luiz Cláudio Navarro, C. Cruz, Francisco Chávez, L. Espada, Patricia Hernández, T. Gallego · 2013
This paper presents an unconventional implementation of an Evolutionary Algorithm with the aim of studying underlying processes influencing human artistic creation. Instead of encoding new software functions that may help Evolutionary Algorithms running on computers to emulate human creativity, we extract the main components of an Evolutionary Algorithm and establish them as a set of rules that must be followed by human artists when creating art. Thus the Evolutionary Algorithm is performed and run by human brains and hands, for a number of generations, applying not only fitness evaluation -as is typically the case with Interactive EAs- but also selection, crossover and mutation. The paper shows how this kind of unplugged EA could become a very useful tool for understanding creative processes, allowing in the future to add new human knowledge in conventional EAs devoted to the creation of art and design.