VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.
Modern Topics in Pattern Recognition
Neuere Themen der Mustererkennung
Last Updated: 2026-02-05 14:55:11
Abstract
In this seminar, recent papers of the pattern recognition and machine learning literature are presented and discussed. Possible topics cover statistical models in computer vision, graphical models and machine learning.
Content
Graphical models are used to specify complex probabilistic models as they are required in real world applications, including computer vision, diagnosis, bioinformatics and machine learning. Belief propagation in graphical models denotes a class of algorithms which are used to efficiently adapt these models to data. Roughly speaking, the question is posed as follows: given a set of variables with statistical dependencies which are represented by a lattice of nodes with interconnecting links (i.e., graphical model), what are the (most probable) states of all the nodes in the lattice when only the states of a (small) group of nodes is known from data? Due to the computational complexity of (exact) belief propagation, it is essential to develop computationally efficient, approximate approaches. In this seminar, we survey state-of-the-art belief-propagation algorithms, and discuss relevant approaches in error-correcting coding and statistical physics that lend itself to belief propagation.
General Information
- Language
- German
- Frequency
- Yearly recurring
Examination
- Type
- session examination
- Mode
- oral 30 minutes
Course Components
| Type | Title | Time & Place | Hours |
|---|---|---|---|
| seminar | Neuere Themen der Mustererkennung |
|
2 h weekly |