Seminar: Blockseminar: Neural ODE and Generative Modelling (Master Seminar) - Details

Seminar: Blockseminar: Neural ODE and Generative Modelling (Master Seminar) - Details

Sie sind nicht in Stud.IP angemeldet.

Allgemeine Informationen

Veranstaltungsname Seminar: Blockseminar: Neural ODE and Generative Modelling (Master Seminar)
Veranstaltungsnummer MTH-1360; MTH-1410
Semester WS 2024/25
Aktuelle Anzahl der Teilnehmenden 4
erwartete Teilnehmendenanzahl 15
Heimat-Einrichtung Mathematische Bildverarbeitung
Veranstaltungstyp Seminar in der Kategorie Lehre
Art/Form Blockseminar
Veranstaltung findet in Präsenz statt / hat Präsenz-Bestandteile Ja
Hauptunterrichtssprache englisch
Weitere Unterrichtssprache(n) Deutsch

Räume und Zeiten

Keine Raumangabe

Modulzuordnungen

Kommentar/Beschreibung

Score-based generative models have demonstrated state-of-the-art performance in numerous applications in recent years. The central concept behind these models involves the gradual injection of noise into the training data, followed by learning the reverse process to generate new samples. The training and sampling procedures can be conducted independently. The learning phase is facilitated by noise-conditional score networks, while sampling can be accomplished through various methods, including Langevin Monte Carlo approaches, stochastic differential equations, ordinary differential equations, and various combinations.

This seminar will overview generative models and common architectures, comparing different training objectives and sampling methods. We will examine normalizing flows, score matching, and Langevin dynamics. Additionally, we will discuss the use of stochastic differential equations (SDEs) in score-based models and conclude with comparisons to other diffusion models and potential enhancements in sample generation.