Wasserstein GANs Work Because They Fail (to Approximate the Wasserstein Distance)

03/02/2021
by   Jan Stanczuk, et al.
0

Wasserstein GANs are based on the idea of minimising the Wasserstein distance between a real and a generated distribution. We provide an in-depth mathematical analysis of differences between the theoretical setup and the reality of training Wasserstein GANs. In this work, we gather both theoretical and empirical evidence that the WGAN loss is not a meaningful approximation of the Wasserstein distance. Moreover, we argue that the Wasserstein distance is not even a desirable loss function for deep generative models, and conclude that the success of Wasserstein GANs can in truth be attributed to a failure to approximate the Wasserstein distance.

READ FULL TEXT

page 4

page 5

page 6

page 7

research
10/02/2019

On the estimation of the Wasserstein distance in generative models

Generative Adversarial Networks (GANs) have been used to model the under...
research
02/17/2023

PAC-Bayesian Generalization Bounds for Adversarial Generative Models

We extend PAC-Bayesian theory to generative models and develop generaliz...
research
12/16/2017

On reproduction of On the regularization of Wasserstein GANs

This report has several purposes. First, our report is written to invest...
research
06/17/2015

Learning with a Wasserstein Loss

Learning to predict multi-label outputs is challenging, but in many prob...
research
02/01/2019

Normalized Wasserstein Distance for Mixture Distributions with Applications in Adversarial Learning and Domain Adaptation

Understanding proper distance measures between distributions is at the c...
research
05/21/2023

Alignment of Density Maps in Wasserstein Distance

In this paper we propose an algorithm for aligning three-dimensional obj...
research
10/28/2019

Approximate Bayesian Computation with the Sliced-Wasserstein Distance

Approximate Bayesian Computation (ABC) is a popular method for approxima...

Please sign up or login with your details

Forgot password? Click here to reset