We introduce new families of Integral Probability Metrics (IPM) for training Generative Adversarial Networks (GAN). Our IPMs are based on matching statistics of distributions embedded in a finite dimensional feature space. Mean and covariance feature matching IPMs allow for stable training of GANs, which we will call McGan. McGan minimizes a meaningful loss between distributions.
@article{arxiv.1702.08398,
title = {McGan: Mean and Covariance Feature Matching GAN},
author = {Youssef Mroueh and Tom Sercu and Vaibhava Goel},
journal= {arXiv preprint arXiv:1702.08398},
year = {2017}
}