Com podem programar l’art? Introducció a les generative adversarial networks

Abstract

[en] Generative art and computational creativity are two areas that have been very popularized between artists, engineers, and theorists of philosophy and psychology in the past few years. Computational advances and their recent applications challenge the creative exclusivity of the human being. But can a computer be creative? We still can’t find an answer and maybe we never will, being art something so personal and subjective, it avoids all sorts of agreement and convention. Nevertheless, we can sow doubt. We will do so by implementing a Generative Adversarial Network, a recent trend in the world of artificial intelligence and deep learning, and training it with the work of a renowned artist. The results will be analyzed, always subjectively, through comparisons between executions, and surveys.