What is involved in the case of artifical neural networks (ANNs) are models that are based on the biological networks in a human brain or the spinal cord.Nerve cells or neurons are connected to one another in the brain in the form of a tightly packed network. Whilst living beings are learning, the connections (synapses) between the cells that are necessary for the learning process are altered. Through this adjustment the brain is able to learn how to cope with different tasks and to solve problems.
This process can be simulated on a computer. Artificial neural networks reproduce the behaviour of the brain which is able to organise itself through learning processes. The primary aim in so doing is to abstract the information processing in itself and not so much gain a technical understanding of biological neural networks and their chemical-physical processes.
As a branch of CI they are simultaneously a branch of artificial intelligence and the research field of neuroinformatics. They can basically be used for any task where it involves identifying connections between “fuzzy” samples. Within the GECO►C research team, ANNs are mainly used in control technology in order to support traditional controllers with forecasts which the network has established from a self-generated forecast with regard to the course of the process.
