Information Technology Reference
In-Depth Information
a.
012345678901234567890123456789012
TDQDbaabababaaaaa1523277252856897-[m]
W = {
-0.202
, -1.934,
-0.17
, 0.013, 1.905, 1.167, 1.801, -1.719, 1.412, 0.434}
m
W = {
1.49
, -1.934,
1.064
, 0.013, 1.905, 1.167, 1.801, -1.719, 1.412, 0.434}
d
012345678901234567890123456789012
TDQDbaabababaaaaa1523277252856897-[d]
b.
NN
m
T
D
Q
D
a
a
a
a
b
b
b
b
NN
d
T
D
Q
D
a
a
a
a
b
b
b
b
Figure 10.4.
Illustration of direct mutation of weights.
a)
The mother and daughter
chromosomes with their respective arrays of weights. In this case, weights at
positions 0 and 2 were mutated. Note that the network architecture is the same for
both mother and daughter.
b)
The mother and daughter neural networks encoded in
the chromosomes. Note that the point mutation at position 2 (weight -0.17) has
manifold effects as this weight appears four different times in the neural network.
Note also that a mutation in the weights at positions 0, 4, 6, and 9 would have had
a neutral effect as these weights have no expression on the neural network.
