Approximating Probability Distributions by Using Wasserstein Generative Adversarial Networks

DOI10.1137/22M149689XarXiv2103.10060OpenAlexW4388295184MaRDI QIDQ6090834FDOQ6090834

Authors: Yihang Gao, Michael Ng 

Publication date: 20 November 2023

Published in: SIAM Journal on Mathematics of Data Science (Search for Journal in Brave)

Abstract: Studied here are Wasserstein generative adversarial networks (WGANs) with GroupSort neural networks as their discriminators. It is shown that the error bound of the approximation for the target distribution depends on the width and depth (capacity) of the generators and discriminators and the number of samples in training. A quantified generalization bound is established for the Wasserstein distance between the generated and target distributions. According to the theoretical results, WGANs have a higher requirement for the capacity of discriminators than that of generators, which is consistent with some existing results. More importantly, the results with overly deep and wide (high-capacity) generators may be worse than those with low-capacity generators if discriminators are insufficiently strong. Numerical results obtained using Swiss roll and MNIST datasets confirm the theoretical results.

Full work available at URL: https://arxiv.org/abs/2103.10060

zbMATH Keywords

capacityapproximationgeneralization errorGroupSort neural networksWasserstein GAN

Mathematics Subject Classification ID

Computational learning theory (68Q32) Analysis of algorithms (68W40)

Cites Work

• Title not available (Why is that?)
• Optimal Transport
• Approximation by superpositions of a sigmoidal function
• Neural Network Learning
• On parameter estimation with the Wasserstein distance
• Community structure in large networks: natural cluster sizes and the absence of large well-defined clusters
• An Iterative Algorithm for Computing the Best Estimate of an Orthogonal Matrix
• Error bounds for approximations with deep ReLU networks
• Central limit theorems for empirical transportation cost in general dimension
• Title not available (Why is that?)
• On the capacity of deep generative networks for approximating distributions

Cited In (5)

• Loss-sensitive generative adversarial networks on Lipschitz densities
• Title not available (Why is that?)
• Quasi-Random Sampling for Multivariate Distributions via Generative Neural Networks
• Inferential Wasserstein Generative Adversarial Networks
• Title not available (Why is that?)

This page was built for publication: Approximating Probability Distributions by Using Wasserstein Generative Adversarial Networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6090834)