Stabilizing GANs with Prediction
Adversarial networks are notoriously hard to train, and simple training methods often collapse. We present a simple modification to the standard training method that increases stability. The method is provably stable for a class of saddle-point problems, and improves performance of numerous GANs.