VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.
Abstract
This course introduces probabilistic deep learning (DL). DL is used for data with complex features like images. We treat DL as probabilistic models, as a continuation of GLMs (logistic regression, ...). The models are fitted with maximum likelihood or Bayesian learning.
Objective
Der Kurs wird auf Deutsch gegeben. Alle Unterrichtsmaterialien sind auf Englisch, daher sind auch die Lernziele auf Englisch formuliert. You will learn about different neural network architectures (e.g. fully connected and convolutional neural networks) and how to choose the appropriate NN architecture for your task at hand. You will learn to model different outcome distributions such as Gaussians, Poissonians, or Multinomial for the task at hand. You will get practical experiences in setting up probabilistic DL models, learn how to tune them, and learn how to control the training procedure.
General Information
- Language
- German
- Levels
- WBZ
- Frequency
- Every two years
Examination
- Type
- ungraded semester performance
Registration & Places
Course Components
| Type | Title | Time & Place | Hours |
|---|---|---|---|
| lecture with exercise |
Deep Learning: Ein probabilistischer Ansatz
Does not take place this semester.
|
No time listed | 19.5 h semesterly |