static void Main(string[] args)
{
FullyConnectedNetwork net = new FullyConnectedNetwork(2,
2, 2, 1);
BackPropagationGradiant teacher = new BackPropagationGradiant(net);
Sample sample1 = new Sample()
{
Inputs = new double[] { 0, 0 },
Outputs = new double[] { 1 },
};
Sample sample2 = new Sample()
{
Inputs = new double[] { 1, 0 },
Outputs = new double[] { -1 },
};
Sample sample3 = new Sample()
{
Inputs = new double[] { 0, 1 },
Outputs = new double[] { -1 },
};
Sample sample4 = new Sample()
{
Inputs = new double[] { 1, 1 },
Outputs = new double[] { 1 },
};
teacher.AddSample(sample1);
teacher.AddSample(sample2);
teacher.AddSample(sample3);
teacher.AddSample(sample4);
Console.WriteLine("*************");
Console.WriteLine("Test XOR gate");
Console.WriteLine("*************");
while (true)
{
for (int x = 0; x < 2; x++)
{
for (int i = 0; i < 20; i++)
{
teacher.LearnOneLoopBase();
}
Console.WriteLine("3 learnings, mean error: " + teacher.MeanQuadraticError);
Console.WriteLine("Ouputs, with {0,1}");
DisplayVector(net.ComputeOutputs(sample2.Inputs));
Console.WriteLine("Ouputs, with {1,1}");
DisplayVector(net.ComputeOutputs(sample4.Inputs));
Console.WriteLine(Environment.NewLine);
}
DisplayNeuronalNetwork(net);
Console.WriteLine("done.");
Console.ReadLine();
Console.Clear();
}
}
