public double[] ActivateNeuralNetwork(double[] environmentOutput)
{
Debug.DLog($"Neural Network Input: {Utilities.ToString(environmentOutput, "f4")}{Utilities.ToString(_turingMachineOutput, "f4")}", true);
// Reset the neural network state
Phenome.ResetState();
// NN Input is the output from the environment, and the output from the turing machine in the previous activation
Phenome.InputSignalArray.CopyFrom(environmentOutput, 0);
Phenome.InputSignalArray.CopyFrom(_turingMachineOutput, environmentOutput.Length);
// Activate the neural network
Phenome.Activate();
double[] nnOutput = new double[Phenome.OutputSignalArray.Length];
Phenome.OutputSignalArray.CopyTo(nnOutput, 0);
Debug.DLog($"Neural Network Output: {Utilities.ToString(nnOutput, "f4")}", true);
// Environment inputs are first of the NN outputs
double[] environmentInput = Utilities.ArrayCopyOfRange(nnOutput, 0, nnOutput.Length - _turingInputLength);
// Turing inputs are the last of the NN outputs
double[] turingMachineInput = Utilities.ArrayCopyOfRange(nnOutput, nnOutput.Length - _turingInputLength, _turingInputLength);
// Activate turing machine with the NN outputs, and store the result for the next iteration, since the environment must be activated in the mean time
_turingMachineOutput = ProcessTuringMachineOutput(TuringMachine.ProcessInput(turingMachineInput));
// Return the environment input (remaining NN outputs)
return(environmentInput);
}
