Back to Search
Start Over
Generalization error of GAN from the discriminator's perspective.
- Source :
- Research in the Mathematical Sciences; 12/28/2021, Vol. 9 Issue 1, p1-31, 31p
- Publication Year :
- 2021
-
Abstract
- The generative adversarial network (GAN) is a well-known model for learning high-dimensional distributions, but the mechanism for its generalization ability is not understood. In particular, GAN is vulnerable to the memorization phenomenon, the eventual convergence to the empirical distribution. We consider a simplified GAN model with the generator replaced by a density and analyze how the discriminator contributes to generalization. We show that with early stopping, the generalization error measured by Wasserstein metric escapes from the curse of dimensionality, despite that in the long term, memorization is inevitable. In addition, we present a hardness of learning result for WGAN. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 25220144
- Volume :
- 9
- Issue :
- 1
- Database :
- Complementary Index
- Journal :
- Research in the Mathematical Sciences
- Publication Type :
- Academic Journal
- Accession number :
- 154372544
- Full Text :
- https://doi.org/10.1007/s40687-021-00306-y