Back to Publications

Author(s) Roziere, B., Teytaud, F., Hosu, V., Lin, H., Rapin, J., Zameshina, M., Teytaud, O.
Title EvolGAN: Evolutionary generative adversarial networks
Abstract We propose to use a quality estimator and evolutionary methods to search the latent space of generative adversarial networks trained on small, difficult datasets, or both. The new method leads to the generation of significantly higher quality images while preserving the original generator’s diversity. Human raters preferred an image from the new version with frequency 83.7% for Cats, 74% for FashionGen, 70.4% for Horses, and 69.2% for Artworks - minor improvements for the already excellent GANs for faces. This approach applies to any quality scorer and GAN generator.
Download RoTeHo20.pdf