Preface
Natural NN's
NN History
  1 2 3 4 5
Artificial NN's
Types of NN's
Programming NN's

The discovery of these problems stopped the wave of research of the neural network. People, who worked with artificial intelligence by the means of conventional programming, sorted out the problems in the network and soon the interest in the neural network collapsed. Only a few incarnated scientists continued the research. Unfortunately, with fewer financial funds.

The research of the neural network went on slowly, and as the computers became more powerful, the neural networks could solve more problems than before. A problem with training hidden neurons was efficiently solved with the introduction of the back propagation algorithm. This algorithm, introduced by Paul Werbos in 1974, gave a neuron permission to let its errors propagate back to former layers of the network. With this form of training, it was possible to solve larger problems. At the same time neural networks became an industrial interest.