Instructors:
Prof. Dr. Stefan Wermter
Displayed in timetable as:
NN - VL
Language of instruction:
English
Min. | Max. participants:
- | 60
Comments/contents:
Neural networks in the brain are the important basis for all behavior and knowledge processing in humans. They produce not only an impressive range of neurocognitive behavior but also an impressive system performance based on computational neural networks. How is this possible and what can we learn from the brain for the development of neural learning and robust computer science systems? This lecture addresses this current and exciting question and provides a comprehensive overview of different computational neural networks and their use and integration into hybrid neural symbolic knowledge-processing systems.
Indicative topics include:
- Neural architectures
- Deep Learning
- Midbrain / cortical architectures
- Neuroscience-inspired robotics
- Neural multimodal integration
- Integration of symbolic, neural or statistical approaches
For details or updates see http://www.informatik.uni-hamburg.de/WTM/teaching/
Learning objectives:
A deeper understanding of artificial neural networks and their integration into computer science system architectures.
Didactic concept:
Lecture with discussions complemented with the associated seminar. As an additional option we can offer access to a neural simulator or a new sophisticated robot simulator in our lab.
Literature:
- Goodfellow I., Bengio Y., Courville A. Deep Learning. MIT Press, 2016.
- Haykin S. Neural networks and learning machines. Prentice Hall, 2008.
- Wermter S., Sun R. Hybrid Neural Systems. Springer Verlag, Heidelberg, 2000.
Additional examination information:
Binding requirement: Lecture and seminar Knowledge Processing with Neural Networks’ can be selected exclusively either as module InfM-NN or as part of the module Integrated Application Subject Neuro-Informatics (IAF NI).
