McGan: Mean and Covariance Feature Matching GAN

Youssef Mroueh; Tom Sercu; Vaibhava Goel;

HIGHLIGHT

We introduce new families of Integral Probability Metrics (IPM) for training Generative Adversarial Networks (GAN).

RELATED

Related Papers     Patents    Grants    Venues    Experts    Cite    Share    Save

ABSTRACT

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.

 Paraphrase     Questionfy     Wikify     Simplify     Summarize