Deep Interactive Evolution

TL;DR

This paper introduces a novel method combining GANs with interactive evolution, enabling controllable, high-quality image creation by evolving latent vectors, demonstrated through a user study on image evolution.

Contribution

It presents a new approach where trained GANs serve as genotype-to-phenotype mappings, allowing evolution of latent vectors for targeted image generation.

Findings

Participants successfully evolved images resembling specific targets.

The method provides controllable image generation with high quality.

GAN-based evolution outperforms traditional interactive methods.

Abstract

This paper describes an approach that combines generative adversarial networks (GANs) with interactive evolutionary computation (IEC). While GANs can be trained to produce lifelike images, they are normally sampled randomly from the learned distribution, providing limited control over the resulting output. On the other hand, interactive evolution has shown promise in creating various artifacts such as images, music and 3D objects, but traditionally relies on a hand-designed evolvable representation of the target domain. The main insight in this paper is that a GAN trained on a specific target domain can act as a compact and robust genotype-to-phenotype mapping (i.e. most produced phenotypes do resemble valid domain artifacts). Once such a GAN is trained, the latent vector given as input to the GAN's generator network can be put under evolutionary control, allowing controllable and high-quality image generation. In this paper, we demonstrate the advantage of this novel approach through a user study in which participants were able to evolve images that strongly resemble specific target images.