Kernel random matrices of large concentrated data: the example of GAN-generated images - Archive ouverte HAL Access content directly
Conference Papers Year :

Kernel random matrices of large concentrated data: the example of GAN-generated images

Abstract

Based on recent random matrix advances in the analysis of kernel methods for classification and clustering, this paper proposes the study of large kernel methods for a wide class of random inputs, i.e., concentrated data, which are more generic than Gaussian mixtures. The concentration assumption is motivated by the fact that one can use generative models to design complex data structures, through Lipschitz-ally transformed concentrated vectors (e.g., Gaussian) which remain concentrated vectors. Applied to spectral clustering, we demonstrate that our theoretical findings closely match the behavior of large kernel matrices, when considering the fed-in data as CNN representations of GAN-generated images (i.e., concentrated vectors by design).
Fichier principal
Vignette du fichier
2019Cicassp.pdf (3.04 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02971224 , version 1 (19-10-2020)

Identifiers

Cite

Mohamed El Amine Seddik, Mohamed Tamaazousti, Romain Couillet. Kernel random matrices of large concentrated data: the example of GAN-generated images. ICASSP 2019 - IEEE International Conference on Acoustics, Speech and Signal Processing, May 2019, Brighton, United Kingdom. ⟨10.1109/ICASSP.2019.8683333⟩. ⟨hal-02971224⟩
66 View
85 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More