On gradient regularizers for MMD GANs
This work addresses training instability in GANs for image generation, offering a more efficient regularization method.
The authors tackled the problem of stabilizing and accelerating training in MMD GANs by proposing a gradient regularizer for the critic, resulting in improved image generation models that outperform state-of-the-art methods on CelebA and ImageNet datasets.
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.