REVIEW 2 cited by
The Numerics of GANs
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
In this paper, we analyze the numerics of common algorithms for training Generative Adversarial Networks (GANs). Using the formalism of smooth two-player games we analyze the associated gradient vector field of GAN training objectives. Our findings suggest that the convergence of current algorithms suffers due to two factors: i) presence of eigenvalues of the Jacobian of the gradient vector field with zero real-part, and ii) eigenvalues with big imaginary part. Using these findings, we design a new algorithm that overcomes some of these limitations and has better convergence properties. Experimentally, we demonstrate its superiority on training common GAN architectures and show convergence on GAN architectures that are known to be notoriously hard to train.
Forward citations
Cited by 2 Pith papers
-
Nested Annealed Training Scheme for Generative Adversarial Networks
NATS, a nested annealed training scheme for GANs, is claimed to improve FID/IS on CIFAR10, LSUN, CelebA, and ImageNet64, but its key theoretical justification is not provided in the preprint.
-
A New Formulation of Lipschitz Constrained With Functional Gradient Learning for GANs
Li-CFG adds an ε-centered gradient penalty to the CFG GAN method and claims this enlarges the discriminator gradient norm, shrinking the latent neighborhood size and thereby increasing image diversity.
Discussion (0). Continue with ORCID to comment.