Solutions to Parallel and Distributed Computing Problems: Lessons from Biological Sciences

10.2 NEURAL-NETWORK DEFINITION

A neural network (Fig. 10.1) is a massively parallel system [7] that stores knowledge and puts it available for use. It is similar to the brain in two aspects: (1) the knowledge is acquired by the network through the learning process; and (2) the connection weights among neurons (synapses) are used to store knowledge. The procedure representing the acquisition of knowledge (i.e., learning algorithm) has the function of modifying the network connection weights to attain the main objective.

Before we proceed we discuss the relationship of neural networks with biological neural systems further [14, 21].

10.2.1 Biological Neural Units

10.2.1.1 Physical (Biological) Neurons

The nervous system consists of two classes of cells: neurons, or nerve cells, and glial, or glial cells. Neurons are the basic building blocks of biological information-processing systems. Glial cells perform more of a support function.

The neurons of the brain can be classified according to their functions. Afferent or sensory neurons provide input to the nervous system; optic nerves are an example. Motor neurons transmit control signals to muscles and glands. Interneuronal neurons process information locally or propagate signals from one site to another, and constitute by far the largest class of cells in the nervous system. There are about 1000 to 10,000 synapses on each neuron. Sensory of chemical stimuli initiates a change in synaptic potential. This is the basis by which one neuron ...