Skip to yearly menu bar Skip to main content


Poster

On gradient regularizers for MMD GANs

Michael Arbel · Danica J. Sutherland · Mikołaj Bińkowski · Arthur Gretton

Room 210 #21

Keywords: [ Kernel Methods ] [ Adversarial Networks ] [ Generative Models ]


Abstract: We propose a principled method for gradient-based regularization of the critic of GAN-like models trained by adversarially optimizing the kernel of a Maximum Mean Discrepancy (MMD). We show that controlling the gradient of the critic is vital to having a sensible loss function, and devise a method to enforce exact, analytical gradient constraints at no additional cost compared to existing approximate techniques based on additive regularizers. The new loss function is provably continuous, and experiments show that it stabilizes and accelerates training, giving image generation models that outperform state-of-the art methods on $160 \times 160$ CelebA and $64 \times 64$ unconditional ImageNet.

Live content is unavailable. Log in and register to view live content