Improving image generative models with human interactions

Andrew Kyle Lampinen, David So, Douglas Eck, Fred Bertsch

Abstract: GANs provide a framework for training generative models which mimic a data distribution. However, in many cases we wish to train a generative model to optimize some auxiliary objective function within the data it generates, such as making more aesthetically pleasing images. In some cases, these objective functions are difficult to evaluate, e.g. they may require human interaction. Here, we develop a system for efficiently training a GAN to increase a generic rate of positive user interactions, for example aesthetic ratings. To do this, we build a model of human behavior in the targeted domain from a relatively small set of interactions, and then use this behavioral model as an auxiliary loss function to improve the generative model. As a proof of concept, we demonstrate that this system is successful at improving positive interaction rates simulated from a variety of objectives, and characterize s

TL;DR: We describe how to improve an image generative model according to a slow- or difficult-to-evaluate objective, such as human feedback, which could have many applications, like making more aesthetic images.

Keywords: human in the loop, GANs, generative adversarial networks, image generative models, computer vision