# On gradient regularizers for MMD GANs

Source: [On gradient regularizers for MMD GANs](<https://www.sota2.com/research/paper/68fa69171e209131cea20245>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-12-22T13:16:46.999Z

Authors: Michael Arbel, Danica J. Sutherland, Miko{\\l}aj Bi\\'nkowski, Arthur Gretton

Published: 2018-05-29T16:23:56.000Z

## 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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1805.11565>)

[PDF](<https://arxiv.org/pdf/1805.11565.pdf>)

[Implementation](<https://www.sota2.com/github.com/MichaelArbel/Scaled-MMD-GAN>)