Backward propagation - a basic algorithm for teaching supervised multi-layer, unidirectional neural networks. It provides a recipe for changing the weight of arbitrary connections of processing elements arranged in neighboring network layers. It is based on the minimization of the sum of the error squares (or other error function) of learning using the optimization method of the largest decrease. By applying a specific method of propagating network learning errors generated at its output, ie sending them from the output layer to the input, the reverse propagation algorithm has become one of the most effective learning algorithms.

wiki

Comments

Popular posts from this blog

List 44

indexation

Mexican pool