Chapter Summary

    Neural networks are one of the most commonly used concepts in Artificial Intelligence. Neural networks are particularly useful for recognizing patterns. They are able to recognize something even when it is distorted.

    A neural network may have input, output, and hidden layers. The input and output layers are the only required layers. The input and output layers may be the same neurons. Neural networks are typically presented input patterns that will produce some output pattern.

    If a neural network mimics the input pattern it was presented with, then that network is said to be autoassociative. For example, if a neural network is presented with the pattern “0110,” and the output is also “0110,” then that network is said to be autoassociative.

    A neural network calculates its output based on the input pattern and the neural network’s internal connection weight matrix. The values for the connection weights will determine the output from the neural network.

    A Hopfield neural network is a fully connected autoassociative neural network. This means that each neuron is connected to every other neuron in the neural network. A Hopfield neural network can be trained to recognize certain patterns. Training a Hopfield neural network involves performing some basic matrix manipulations on the input pattern that is to be recognized.

    This chapter explained how to construct a simple Hopfield neural network. This network was trained using a simple training algorithm. Training algorithms allow weights to be adjusted to produce the desired outputs. There are many advanced training algorithms. Chapter 4 will introduce more complex training algorithms.