100 step rule
The 100-step rule is used in neuroinformatics to demonstrate the performance of neural networks and to motivate massive parallel processing in the connectionism model .
It says:
- “A person can recognize an object or a known person within 0.1 s. With an assumed processing time of a nerve cell of 1 ms, a maximum of 100 sequential processing steps are necessary in the human brain. "
However, no information is given about the total number of processing steps, since the detection takes place massively in parallel.
Today's computers are generally not capable of comparable performance. This statement shows the possibilities of parallel processing and is a justification for the interest of neuroinformatics in biologically motivated processes such as artificial neural networks .
literature
- Andreas Zell: Simulation of neural networks , R. Oldenbourg Verlag, Munich 1997, ISBN 3-486-24350-0 .