Intervention Generative Adversarial Networks

TL;DR

This paper introduces Intervention GANs (IVGAN), a new method that stabilizes GAN training and reduces mode collapse by adding an intervention loss, leading to more reliable and diverse image generation.

Contribution

The paper proposes a novel intervention loss for GANs, improving training stability and diversity, supported by theoretical analysis and extensive experiments.

Findings

Enhanced training stability of GANs.

Reduced mode collapse in image generation.

Effective on multiple datasets including MNIST.

Abstract

In this paper we propose a novel approach for stabilizing the training process of Generative Adversarial Networks as well as alleviating the mode collapse problem. The main idea is to introduce a regularization term that we call intervention loss into the objective. We refer to the resulting generative model as Intervention Generative Adversarial Networks (IVGAN). By perturbing the latent representations of real images obtained from an auxiliary encoder network with Gaussian invariant interventions and penalizing the dissimilarity of the distributions of the resulting generated images, the intervention loss provides more informative gradient for the generator, significantly improving GAN's training stability. We demonstrate the effectiveness and efficiency of our methods via solid theoretical analysis and thorough evaluation on standard real-world datasets as well as the stacked MNIST dataset.