Random Matrix Theory Proves that Deep Learning Representations of GAN-data Behave as Gaussian Mixtures
Résumé
This paper shows that deep learning (DL) representations of data produced by generative ad-versarial nets (GANs) are random vectors which fall within the class of so-called concentrated random vectors. Further exploiting the fact that Gram matrices, of the type G = X X with X = [x 1 ,. .. , x n ] ∈ R p×n and x i independent concentrated random vectors from a mixture model, behave asymptotically (as n, p → ∞) as if the x i were drawn from a Gaussian mixture, suggests that DL representations of GAN-data can be fully described by their first two statistical moments for a wide range of standard classifiers. Our theoretical findings are validated by generating images with the BigGAN model and across different popular deep representation networks.
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...