Probabilistic Graphical Models
(Formerly Machine Learning: Statistical Approaches II)
Winter Semester 2014/2015
The course covers advanced topics in machine learning, for example: Graphical models in machine learning, inference mechanisms and sampling strategies in graphical models, Gaussian processes, probabilistic topic models, unsupervised and semi-supervised learning.
Quick links
Organization
| Module no. | 20-00-0449-iv |
| Event type | integrated course (IV4, 6CP, 4SWS), elective |
| Being taught | regularly, usually winter semesters |
| Time | Wednesdays, 09:50 – 13:20 |
| First class | 15.10.2014 |
| Location | S3|05, room 073 |
| Lecturer | Prof. Stefan Roth, Ph.D. |
| Exam | Usually individual oral exams. |
Content
- Refresher of probability & Bayesian decision theory
- Directed and undirected models and their properties
- Inference in tree graphs
- Approximate inference in general graphs: Message passing and mean field
- Learning of directed and undirected models
- Sampling methods for learning and inference
- Modeling in example applications, including topic models
- Deep networks
- Semi-supervised learning
Learning goals
After successfully attending the course, students have developed an in-depth understanding of probabilistic graphical models. They describe and analyze properties of graphical models, and formulate suitable models for concrete estimation and learning tasks. They understand inference algorithms, judge their suitability and apply them to graphical models in relevant applications. Moreover, they determine which learning algorithms are suitable to estimate the model parameters from example data, and apply these.
Degree courses
Can be taken for credit toward BSc / MSc Informatik, MSc Visual Computing, MSc Autonome Systeme, BSc / MSc Computational Engineering and others. Students from other departments, e.g. Mathematik, Elektrotechnik, IST, or Physik are welcome, though academic credit may need to be arranged.
Prerequisites
It is recommended to have taken Statistical Machine Learning (formerly Machine Learning: Statistical Approaches I).
Literature
|