CAN: Creative Adversarial Networks, Generating "Art" by Learning About Styles and Deviating from Style Norms

We propose a new system for generating art. The system generates art by at art and learning about style; and becomes creative by increasing the potential of the generated art by deviating from the learned styles. We over Generative Adversarial Networks (GAN), which have shown the ability learn to generate novel images simulating a given distribution. We argue such networks are limited in their ability to generate creative products their original design. We propose modifications to its objective to make it of generating creative art by maximizing deviation from established and minimizing deviation from art distribution. We conducted experiments compare the response of human subjects to the generated art with their to art created by artists. The results show that human subjects could distinguish art generated by the proposed system from art generated by artists and shown in top art fairs. Human subjects even rated the images higher on various scales.