Exploring expression-based generative adversarial networks

Abstract

Before the advent of Generative Adversarial Networks (GANs) in 2014, Evolutionary Computation approaches made up the majority of the state of art for image generation. Such approaches have been replaced with adversarial models using deep convolutional networks specifically tailored for GPU computing. Nevertheless, motivated by recent successes in GPU-accelerated Genetic Programming (GP) and given the well-established disposition of expression-based solutions towards image evolution, the prospect of bridging the gap between using symbolic expressions as generators for GANs, instead of neural networks, seems enticing.