Intervention Generative Adversarial Networks
This addresses training challenges in GANs for generative modeling, but it is incremental as it builds on existing GAN frameworks.
The paper tackles the instability and mode collapse problems in Generative Adversarial Networks (GANs) by introducing an intervention loss regularization term, resulting in improved training stability as demonstrated through theoretical analysis and evaluations on datasets like stacked MNIST.
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.