Secured Wireless Network Based on a Novel Dual Integrated Neural Network Architecture
Algorithm 1
DINN.
Input: neural network α, neural network β, loss function ꭈ, learning parameter Ѓ Total number of epochs È
Output: neural network β
(1)
Start
(2)
Neural network β = weight-Initialization
(3)
for to do
(4)
Rearrangement of the Dataset
(5)
for to | Dataset | do
(6)
Extract the xth sample from Dataset
(7)
Forward propagation of the sample by the neural network α to determine the output probabilities evaluate the loss through the output probability Backpropagate the loss through the neural network β.
(8)
Weight-updating of the neural network β by the learning parameter Ѓ.
