Loading...
conference paper
Reducing Noise in GAN Training with Variance Reduced Extragradient
2019
Advances In Neural Information Processing Systems 32 (Nips 2019), 32
We study the effect of the stochastic gradient noise on the training of generative adversarial networks (GANs) and show that it can prevent the convergence of standard game optimization methods, while the batch version converges. We address this issue with a novel stochastic variance-reduced extragradient (SVRE) optimization algorithm, which for a large class of games improves upon the previous convergence rates proposed in the literature. We observe empirically that SVRE performs similarly to a batch method on MNIST while being computationally cheaper, and that SVRE yields more stable GAN training on standard datasets.
Type
conference paper
Authors
Publication date
2019
Published in
Advances In Neural Information Processing Systems 32 (Nips 2019), 32
Peer reviewed
REVIEWED
EPFL units
Event name | Event place | Event date |
Vancouver, CANADA | Dec 08-14, 2019 | |
Available on Infoscience
February 18, 2020
Use this identifier to reference this record