Wasserstein Proximal of GANs
This addresses training instability and inefficiency in GANs for generative modeling, representing an incremental improvement with a novel regularization approach.
The paper tackles the problem of training generative adversarial networks (GANs) by introducing a method based on Wasserstein-2 metric proximal on generators, which improves training speed and stability, as demonstrated by reductions in wall-clock time and Fréchet Inception Distance in experiments.
We introduce a new method for training generative adversarial networks by applying the Wasserstein-2 metric proximal on the generators. The approach is based on Wasserstein information geometry. It defines a parametrization invariant natural gradient by pulling back optimal transport structures from probability space to parameter space. We obtain easy-to-implement iterative regularizers for the parameter updates of implicit deep generative models. Our experiments demonstrate that this method improves the speed and stability of training in terms of wall-clock time and Fréchet Inception Distance.