Neural Networks: A General form of Non-Linear Extrapolation
In many tasks, computers are much faster than humans. However, there are still tasks, such as handwriting recognition, that we humans do much better and much faster than even the fastest modern computers. To help computers solve these tasks, it is natural to simulate the way our brain (which is, in essence, a neural network) solves them. The resulting universal extrapolation tool, called artificial neural network, is described in this lesson. For these networks, continuous mathematics helps to choose the optimal parameters.
Unable to display preview. Download preview PDF.