We investigate the training and performance of generative adversarial networks using the Maximum Mean Discrepancy (MMD) as critic, termed MMD GANs. As our main theoretical contribution, we clarify the situation with bias in GAN loss functions raised by recent work: we show that gradient estimators used in the optimization process for both MMD GANs and Wasserstein GANs are unbiased, while gradients of the generator's theoretical loss are biased in both cases. We discuss the issue of kernel choice for the MMD critic, and characterize the kernel corresponding to the energy distance used for the Cram\'er GAN critic. Being an integral probability metric, the MMD benefits from training strategies recently developed for Wasserstein GANs. In experiments, the MMD GAN is able to employ a smaller critic network than the Wasserstein GAN, resulting in a simpler and faster-training algorithm with matching performance. We also propose an improved measure of GAN convergence, the Kernel Inception Distance, and show how to use it to dynamically adapt learning rates during GAN training.

通过使用最大均值差异（MMD）作为批判家来研究生成对抗网络的训练和性能，我们澄清了最近的工作中提出的GAN损失函数中的偏差问题，并讨论了MMD批判家的内核选择问题，提出了改进的GAN收敛度量，核启动距离，并展示如何在GAN训练期间使用它动态调整学习速率。

解密 MMD GANs