System and method for implementing a neural network

A method of implementing a neural network having an input layer made up of a plurality of input layer nodes, an output layer made up of a plurality of output layer nodes, and a plurality of hidden layers connected between the input layer and the output layer, each hidden layer including a plurality of hidden layer nodes, wherein the input, hidden, and output nodes include processing elements for processing data received from nodes of a previous layer, the data flowing in a forward direction from an input to an output of the neural network sequentially through the input layer, the hidden layers, and the output layer, wherein output data is compared with a target represented by an output label, and wherein a result of the comparison between the output data and the output label is used to modify hidden nodes, comprising the steps of:

• instead of modifying the hidden nodes by backpropagation from the output to the input through the hidden layers in a reverse sequence, modifying the hidden layers by:
• (a) supplying input data, said output label and internal teaching labels to the neural network,
• (b) causing the neural network to process the input data through the hidden layers and outputting a result of the processing for comparison with the output label,
• (c) supplying the internal teaching labels to the hidden layers and calculating scores for the hidden layers based on the internal teaching labels,
• (d) modifying the hidden layers based on the calculated scores and the comparison of the processing result with the output label.