Neural network: Difference between revisions

Latest revision as of 14:44, 20 February 2012

neural network

Usually used to mean "artificial neural network," a computer program inspired by simple models of the brain.

It consists of a network of nodes connected by weighted links that establish the relationships between nodes. Each node sums the weighted inputs entering it and compares the result to a (usually) nonlinear function to produce its own output. Most neural networks have a training rule establishing how the weights are adjusted to bring the average output closest to that desired. The term can also mean the computer chip containing the fully trained (unmodifiable) system used in automatic sensors and controls.
See also machine learning.

INDEX
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
0
1
2
3
4
5
6
7
8
9

@@ Line 9: / Line 9: @@
    </div>
-<div class="definition"><div class="short_definition">Usually used to mean &ldquo;artificial neural network,&rdquo; a computer program inspired by  simple models of the brain.</div><br/> <div class="paragraph">It consists of a network of nodes connected by weighted links that establish the relationships  between nodes. Each node sums the weighted inputs entering it and compares the result to a  (usually) [[nonlinear]] function to produce its own [[output]]. Most neural networks have a training  [[rule]] establishing how the weights are adjusted to bring the average output closest to that desired.  The term can also mean the computer chip containing the fully trained (unmodifiable) system  used in automatic sensors and controls. <br/>''See also'' [[machine learning]].</div><br/> </div>
+<div class="definition"><div class="short_definition">Usually used to mean "artificial neural network," a computer program inspired by  simple models of the brain.</div><br/> <div class="paragraph">It consists of a network of nodes connected by weighted links that establish the relationships  between nodes. Each node sums the weighted inputs entering it and compares the result to a  (usually) [[nonlinear]] function to produce its own [[output]]. Most neural networks have a training  [[rule]] establishing how the weights are adjusted to bring the average output closest to that desired.  The term can also mean the computer chip containing the fully trained (unmodifiable) system  used in automatic sensors and controls. <br/>''See also'' [[machine learning]].</div><br/> </div>
 </div>