Users Also Read
MCQ's Search Engine
Electrical Engineering
Mechanical Engineering
Civil Engineering
Automobile Engineering
Chemical Engineering
Computer Engineering
Electronics Engineering
Medical Science Engg
All Engineering Dictionary Terms
Definition of "pruning self-generating neural network" |
a methodology for reducing the size of self generating neural network (SGNN). During SGGN training, the network may grow very quickly, and some parts of the network may become useful in neither training nor classification, and the weights of these parts of the neurons of the network never change after some training stage. We call these parts of the network dead subnets. It is obvious that the dead subnets of the network should be pruned away to reduce the network size and improve the network performance. One way to achieve this is to check the weights of each neuron in the network to see whether they have been changed since last training epoch (or during the last few epochs). If they are unchanged, the neuron may be evidently dead and should be removed from the network. If a neuron is removed from an SGNN, all of its offspring should also be removed. |
Please type any word or choose alphabet below... |
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 1 2 3 4 5 6 7 8 9 0 |