VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.
Diffusion Models, Sampling and Stochastic Localization
Last Updated: 2026-02-05 16:37:23
Abstract
This is a master/PhD level course to prepare students for academic research in Markov chain Monte Carlo sampling algorithms, diffusion generative models and the associated theory. We cover basic sampling algorithms, provide geometric intuitions and show frameworks for proving results in log-concave sampling.
Objective
The main objective is to learn about the basic sampling algorithms such as ball walk, Langevin algorithms and Hamiltonian Monte Carlo. From these classic sampling algorithms, we make connections with the diffusion generative models, which are the state-of-the-art models for generating natural images. Additionally, we cover theoretical frameworks to prove convergence guarantees of sampling algorithms, including the stochastic localization technique.
Resources
Learning Materials (Links)
- Main link
- Information
General Information
- Language
- English
- Levels
- MSC
Examination
- Type
- ungraded semester performance
Course Components
| Type | Title | Time & Place | Hours |
|---|---|---|---|
| lecture | Diffusion Models, Sampling and Stochastic Localization |
|
14 h semesterly |
Offered In
-
-
Electives (For the Master's degree in Applied Mathematics the following additional condition (not manifest in myStudies) must be obeyed: At least 14 of the required 26 credits from core courses and electives must be acquired in areas of applied mathematics and further application-oriented fields.)
-
-
-
Statistics Master (The following courses belong to the curriculum of the Master's Programme in Statistics. The corresponding credits do not count as external credits even for course units where an enrolment at ETH Zurich is not possible.)