Canadian scientists at the University of Waterloo’s Centre for Theoretical Neuroscience have built a functioning computer simulation of the human brain, called Spaun. While human brain has 80-100 billion neurons, the artificial brain contains 2.5 million computer-simulated neurons. Spaun stands for Semantic Pointer Architecture Unified Network. It is a computer model that can recognize and remember numbers, figure out numeric sequences, and write them down with a simulated arm. It has an eye that can view a 28-by-28-pixel image.
Chris Eliasmith, et. al
The simulated neurons in Spaun model are designed to behave like human neurons as much as possible. The subsystems in this model resemble brain regions including prefrontal cortex, basil ganglia and thalamus. The model performs eight diverse tasks in order to capture several aspects of neuroanatomy, neurophysiology, and psychological behavior. The tasks include recognizing a digit or adding two digits shown in succession, identifying the position of a digit in a list and making basic inferences like completing a series of numbers.
Compared to human neural network, Spaun is relatively simple, hard-wired, lacking the plasticity and adaptive capability. Still this invention is very crucial as it will be useful to further the research in the areas of artificial brain.
