Integrated Seminar: Neural Networks / Neuronale Netzwerke

News/Aktuelles

• Important: We will run the seminar as a block in the end of the semester and will NOT meet weekly in class. The first meeting for the module (lecture + seminar) will be on 07.04.2016, 10:15h, room D-220
• For this course we use the MIN-CommSy.
• First registration period/Erste Anmeldephase (STiNE): 15.02.16 09:00h - 03.03.16 13:00h

Information/Allgemeine Informationen
LV-Nummer:	64-417
Lecturer:	Dr. Cornelius Weber, Prof. Stefan Wermter
Period:	Will be run as a block in the end of the lecture period
Room:	Will be announced.
Credit Hours	2 SWS
Language:	English with english material
Module:	InfM-NI2, InfM-NN, IAF-NI2, MV-ISR2-WV2, InfM-WV2

Contents/Inhalte

This seminar provides a focused overview of artificial neural network approaches and their use and integration into real-world knowledge-processing systems.
Indicative topics include

• Neuroscience-inspired Architectures
• Neural architectures
• Spiking neural networks
• Midbrain / cortical architectures
• Neuroscience-inspired robotics
• Neural multimodal integration
• Integration of symbolic, neural or statistical approaches

Dieses Seminar liefert einen fokussierten überblick über neue Ansätze künstlicher neuronaler Netzwerke und deren Verwendung und Integration in reale wissensbasierte Systeme.
Indikative Themen sind:

• Neuronale Architekturen
• Spikende neuronale Netze
• Subkortikale und Kortikale Architekturen
• Neuro-inspirierte Robotik
• Multimodale neuronale Integration
• Integration von symbolischen, neuronalen und statistischen Ansätzen

Objective/Lernziel

Critically analyse and evaluate neural network models and scientific literature and present research in the area of knowledge processing with neural architectures.

Die Fähigkeit zur kritischen Analyse von neuronalen Modellen und der Präsentation von Forschungsthemen im Bereich Wissensverarbeitung mit neuronalen Netzwerken.

Procedure/Vorgehen

Seminar with discussions. Students can chose either literature study or practical study of implemented neural models as a basis for their paper and presentation. For the practical choice, we will offer access to neural simulation enviroments or a new sophisticated robot simulator in our lab.

Seminar mit Diskussionen. Ausarbeitung und Vortrag beruhen wahlweise entweder auf Literaturstudium oder auf praktischer Implementation/Exploration von neuronalen Modellen. Wir stellen dazu eine Auswahl von neuronalen Simulationsumgebungen oder eine neue Robotersimulationsumgebung zur Verfügung.

Literature/Literatur

Will be suggested in the first seminar. Some indicative initial reference is listed below
/Wird am ersten Termin besprochen. Einige indikative Vorschläge sind:

• Wermter S., M. Page, M. Knowles, V. Gallese, F. Pulvermuller and J. Taylor. Neural Network: Special Issue on What it Means to Communicate. Volume 22, Number 2, March, 2009.
• Wermter S., Palm G., Elshaw M. (Eds.) Biomimetic Neural Learning for Intelligent Robots. Springer, Heidelberg, Germany. 2005.
• Wermter S., Austin J., Willshaw D. (Eds.) Emergent Neural Computational Architectures based on Neuroscience. Springer, Heidelberg, Germany, 2001.

Material

Templates for seminar papers and presentations as well as further hints can be found on the web page: Hints for Seminars
The material for this course can be found within the MIN-CommSy portal:
Informatik-CommSy: WTM: Neural Networks SS2016