Jordan network without an unnoticed neuron layer for our training attempts to ensure the output neurons may directly provide input. This method is a strong simplification because generally more complicated networks are used. But yet this does not alter the learning principle.