Knowledge-Based Systems

7.1 introduCtion 191

The major disadvantage of an artificial neural network is that it can-

not fully mimic the human brain functionalities or intelligence. A neural

network cannot offer facilities like explicit explanation and reasoning; as

knowledge is stored implicitly in generalized connections between neurons.

Moreover, scaling up a neural network is not a straightforward affair. Sup-

pose we trained a neural network for, say, 200 input neurons and now we

wish to extend this to a neural network of 201 input neurons. Typically, we

have to begin again with an entire training session for the new network.

7.1.2 Comparing Artificial Neural Networks with the 

von Neumann Model

In this subsection, we will look at some featur ...