Regularized Generative Adversarial Network
This work addresses the practical problem of generating samples from a different distribution than the training data, which is relevant for artists using generative methods.
The paper introduces Regularized Generative Adversarial Network (RegGAN), a framework designed to generate samples from a target probability distribution that is distinct from the training data distribution. This is achieved by simultaneously training a generator and two discriminators within an adversarial process.
We propose a framework for generating samples from a probability distribution that differs from the probability distribution of the training set. We use an adversarial process that simultaneously trains three networks, a generator and two discriminators. We refer to this new model as regularized generative adversarial network (RegGAN). We evaluate RegGAN on a synthetic dataset composed of gray scale images and we further show that it can be used to learn some pre-specified notions in topology (basic topology properties). The work is motivated by practical problems encountered while using generative methods in the art world.