Question: Apply the extended Hebb network procedure, written in the previous exercise to realize the Boolean functions AND and OR, as illustrated in Figure. What happens when the first, second, and third patterns are presented to the network in the AND function? The same happens with the first pattern of function OR. Explain why the learning rule is not capable of generating a correct decision line? Change the binary vector of desired outputs by a bipolar vector, that is, every value ‘0' in vector d should correspond to a value ‘-1', as illustrated in Figure, and reapply the Hebb network to this problem. What happens? Is the network capable of solving the problem? Justify your answer. Finally, change the input patterns into the bipolar form and try to solve the problem using the simple Hebb network.
![612_4.47.png](https://secure.tutorsglobe.com/CMSImages/612_4.47.png)