This course provides an introduction to
Computational Neuroscience, and also into related engineering
disciplines (large scale simulation of brain systems, neuromorphic
engineering). Furthermore we will discuss efforts by major
companies such as IBM to change paradigms for the design of
hardware and software in digital computer. In particular some of
the content of the new book "Smart Machines: IBM's Watson and the
Era of Cognitive Computing" will be discussed (a copy of the book
is available for the students of this course on the course webpage
" Literature")
This course is independent from the course "Neuronale Netzwerke
A", and does not require knowledge from it. But it requires
knowledge of the basic concepts related to neural networks that
are presented in the undergraduate courses Computational
Intelligence or Einführung in die Wissensverarbeitung). That
material is contained in the scriptum
https://igi-web.tugraz.at/lehre/CI/CI-skript.pdf
No prior knowledge from biology or brain science is assumed.
Computer science is not only the science of digital computing
machines, but also the science of computation and information
processing in biological systems, e.g. in the brain. In fact, the
brain is at present still the best performing (and most energy
efficient) information processing systems, hence there are good
chances that computer science may profit from further insight into
information processing in the brain. This is in fact one of the
goals of the 10-year EU Flagship Project "Human Brain Project"
https://www.humanbrainproject.eu/
whose research strategies will be discussed in this course.
Our institute is responsible for the Work Package "Principles of
Brain Computation" in this project, which started in October 2013.
This course will present the best current models for biological
neurons, synapses, and concepts for understanding information
processing in networks of biological neurons. We will discuss
several competing hypotheses regarding the organization of
information processing in the brain, in particular a rather new
one where one views the brain as a probabilistic inference (and
-learning) machine. Hence we will also provide a short
self-contained introduction into probabilistic inference, which
has turned out to be an essential tool for modern Artificial
Intelligence and Machine Learning.
In the practical exercises, the students learn to implement
several of these models with state-of-the-art software systems,
and can experiment with them on their own.