Generalized normalizing flows via Markov chains /

Normalizing flows, diffusion normalizing flows and variational autoencoders are powerful generative models. This Element provides a unified framework to handle these approaches via Markov chains. The authors consider stochastic normalizing flows as a pair of Markov chains fulfilling some properties,...

Ausführliche Beschreibung

Gespeichert in:
Hauptverfasser:
Hagemann, Paul Lyonel, Hertrich, Johannes, Steidl, Gabriele
Format:
Elektronisch E-Book
Sprache:
Englisch
Veröffentlicht:
Cambridge : Cambridge University Press, 2023.
Ausgabe:
First edition.
Zusammenfassung:
Normalizing flows, diffusion normalizing flows and variational autoencoders are powerful generative models. This Element provides a unified framework to handle these approaches via Markov chains. The authors consider stochastic normalizing flows as a pair of Markov chains fulfilling some properties, and show how many state-of-the-art models for data generation fit into this framework. Indeed numerical simulations show that including stochastic layers improves the expressivity of the network and allows for generating multimodal distributions from unimodal ones. The Markov chains point of view enables the coupling of both deterministic layers as invertible neural networks and stochastic layers as Metropolis-Hasting layers, Langevin layers, variational autoencoders and diffusion normalizing flows in a mathematically sound way. The authors' framework establishes a useful mathematical tool to combine the various approaches.
Umfang:
1 online resource (57 pages) : illustrations (black and white, and colour), digital, file(s).
Anmerkungen:
Previously issued in print: 2022.
Zielpublikum:
Specialized.
Bibliografie:
Includes bibliographical references.
ISBN:
9781009330992
1009330993
9781009331036
1009331035
9781009331012
1009331019
Schlagworte:
Markov processes.
Bezugswerke:
Links: