Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Vardanyan, Elen, Hunanyan, Sona, Galstyan, Tigran, Minasyan, Arshak, Dalalyan, Arnak
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914825943121920
author Vardanyan, Elen
Hunanyan, Sona
Galstyan, Tigran
Minasyan, Arshak
Dalalyan, Arnak
author_facet Vardanyan, Elen
Hunanyan, Sona
Galstyan, Tigran
Minasyan, Arshak
Dalalyan, Arnak
contents This paper explores the problem of generative modeling, aiming to simulate diverse examples from an unknown distribution based on observed examples. While recent studies have focused on quantifying the statistical precision of popular algorithms, there is a lack of mathematical evaluation regarding the non-replication of observed examples and the creativity of the generative model. We present theoretical insights into this aspect, demonstrating that the Wasserstein GAN, constrained to left-invertible push-forward maps, generates distributions that avoid replication and significantly deviate from the empirical distribution. Importantly, we show that left-invertibility achieves this without compromising the statistical optimality of the resulting generator. Our most important contribution provides a finite-sample lower bound on the Wasserstein-1 distance between the generative distribution and the empirical one. We also establish a finite-sample upper bound on the distance between the generative distribution and the true data-generating one. Both bounds are explicit and show the impact of key parameters such as sample size, dimensions of the ambient and latent spaces, noise level, and smoothness measured by the Lipschitz constant.
format Preprint
id arxiv_https___arxiv_org_abs_2307_16422
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution
Vardanyan, Elen
Hunanyan, Sona
Galstyan, Tigran
Minasyan, Arshak
Dalalyan, Arnak
Statistics Theory
Machine Learning
This paper explores the problem of generative modeling, aiming to simulate diverse examples from an unknown distribution based on observed examples. While recent studies have focused on quantifying the statistical precision of popular algorithms, there is a lack of mathematical evaluation regarding the non-replication of observed examples and the creativity of the generative model. We present theoretical insights into this aspect, demonstrating that the Wasserstein GAN, constrained to left-invertible push-forward maps, generates distributions that avoid replication and significantly deviate from the empirical distribution. Importantly, we show that left-invertibility achieves this without compromising the statistical optimality of the resulting generator. Our most important contribution provides a finite-sample lower bound on the Wasserstein-1 distance between the generative distribution and the empirical one. We also establish a finite-sample upper bound on the distance between the generative distribution and the true data-generating one. Both bounds are explicit and show the impact of key parameters such as sample size, dimensions of the ambient and latent spaces, noise level, and smoothness measured by the Lipschitz constant.
title Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution
topic Statistics Theory
Machine Learning
url https://arxiv.org/abs/2307.16422