Some theoretical properties of GANs

DOI10.1214/19-AOS1858zbMath1451.62107arXiv1803.07819OpenAlexW3043551088MaRDI QIDQ2196234

Maxime Sangnier, Ugo Tanielian, Gérard Biau, Benoît Cadre

Publication date: 28 August 2020

Published in: The Annals of Statistics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1803.07819

zbMATH Keywords

neural networkscentral limit theoremgenerative modelsJensen-Shannon divergenceadversarial principle

Mathematics Subject Classification ID

Asymptotic properties of parametric estimators (62F12) Central limit and other weak theorems (60F05) 2-person games (91A05) General topics in artificial intelligence (68T01) Neural nets and related approaches to inference from stochastic processes (62M45)

Related Items (6)

Monte Carlo simulation of SDEs using GANs ⋮ A convenient infinite dimensional framework for generative adversarial learning ⋮ Unnamed Item ⋮ Minimax rate of distribution estimation on unknown submanifolds under adversarial losses ⋮ Posterior asymptotics in Wasserstein metrics on the real line ⋮ Finite sample properties of parametric MMD estimation: robustness to misspecification and dependence

Uses Software

• TensorFlow
• VEEGAN
• Wasserstein GAN
• f-GAN

Cites Work

• Unnamed Item
• Universal smoothing factor selection in density estimation: theory and practice. (With discussion)
• A new metric for probability distributions

This page was built for publication: Some theoretical properties of GANs