Computational Intelligence

This subject area catalog gives an introduction to the most important currently known methods to make machines "intelligent". The courses introduce the most important methods and concepts of machine learning, adaptive robots, neural networks, computational neuroscience, as well as speech and signal processing. In accompanying practice classes the methods are applied on different real world examples to deepen the knowledge and their usage.

The old vision of building machines that can think and learn has turned out to be much more difficult than expected a few decades ago. Nevertheless, significant progress in important subareas and applications has been made, which nowadays increasingly influences the development of new products. The relevant methods are presented in the classes of this catalog. Understanding the theoretical concepts and carrying out experiments with state-of-the-art software goes hand in hand.

Research on computation and learning in the nervous systems of biological organisms has not only inspired neural networks, but a lot of other machine learning methods and lead to the development of new paradigms for processing complex sensory data in real-time. Within this catalog, interested students can gain insight to current research topics in this exciting field, where artificial intelligence and computational neuroscience overlap. Experience is gained both in theoretical modelling and in computer simulations of neural computation and learning in living organisms.
 
 

Courses

Note:

The courses Computational Intelligence and Signal Processing are compulsory for the Telematics bachelor program. All further courses in this catalog therefore assume that students are familiar with the material covered in these lectures, especially in the areas of machine learning, neural networks, signal processing, statistics and probability calculus.

All the classes are offered in English (on demand). Therefore this course package is very suitable for international students, too.

The catalog consists of the following courses:

Title	VO	UE
Adaptive Systems	2	1
Advanced Signal Processing 1	0	2
Machine Learning A *	2	1
Machine Learning B **	2	1
Neural Networks A *	2	1
Neural Networks B	2	1
Nonlinear Signal Processing *	2	1
Speech Communication 2	2	0
Speech Communication Laboratory	0	2
Computational Intelligence Seminar A-F (6 Seminars)	0	2 each
Seminar/Projekt Machine Learning and Neuroinformatics	0	6
Seminar/Projekt Signal Processing	0	6

* Shaded areas identify the core courses which have to be selected by all students following this program.
** Since Machine Learning A is not offered in WS 2006/07, students may instead attend Machine Learning B as a core course.

All of the courses can be attended independently of the others (e.g. Neural Networks A is not a prerequisite for Neural Networks B). The lectures are offered in a 2-year cycle, so students can complete the whole catalog and finish their diploma thesis within 2 or 3 years. It is also possible to design an individual catalog, which best matches your interests and takes into account actually offered courses.
If you are interested in the Computational Intelligence catalog, or if you would like to arrange an individual catalog, please contact in advance any of the persons listed here.  
 

Characteristics of the Courses

Lectures and design practicals can not be passed only with passive learning and a final exam. Theoretical and practical exercises enable an interactive approach to the best possible understanding of the contents.

In our seminars students gain experience with the most recent research results and learn how to apply this knowledge to concrete problems.