MR-GAN: Manifold Regularized Generative Adversarial Networks for Scientific Data

Despite the growing interest in applying generative adversarial networks (GANs) in complex scientific applications, training GANs on scientific data remains a challenging problem from both theoreti...