Improving image generative models with human interactions

TL;DR

This paper presents a method to enhance GANs by incorporating a human behavior model as an auxiliary loss, enabling the generation of images that increase positive user interactions, demonstrated through simulations.

Contribution

The authors develop a system that uses a learned human behavior model as an auxiliary loss to improve GANs for human-interaction objectives.

Findings

Successfully improves positive interaction rates in simulated data

Effectiveness depends on the accuracy of the behavioral model

Characterizes factors influencing system performance

Abstract

GANs provide a framework for training generative models which mimic a data distribution. However, in many cases we wish to train these generative models 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 improving a GAN to target an objective involving human interaction, specifically generating images that increase rates of positive user interactions. To improve the generative model, 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. We show that this system is successful at improving positive interaction rates, at least on simulated data, and characterize some of the factors that affect its performance.