Information Technology Reference
In-Depth Information
Fig. 4.2 Natural (a) and
“formal” (b, c) neural
networks
(Fig.
4.2
). Sensory neurons were considered sensitive elements which generate and
send a signal to the network under the influence of some external stimulus (such as
electric signals, light, sound, etc.). Associative neurons were defined as logic
elements that issue an output signal when the algebraic sum of input signals exceeds
a certain threshold value. Reactive neurons are elements that receive input from
sensory neurons and form control stimuli in the external environment. Rosenblatt
developed an operating device—single-layer perceptron—capable of classifying
input signals into one of two classes.
Rosenblatt's approach was used in the early 1960s to explain various psycho-
logical and physiological phenomena. But, in general, interest in them soon
dropped significantly because of the absence of demand for neural network ideas
