Question

neural network

thank you

4. Implement an XOR network for two state variables; calculate the synaptic weights for first three runs.

Answer 1

is equivalent to... AB AB AB AB

Let the input of XOR gate be ‘a’ and ‘b’. Let the output be ‘Y’.
So, we know that the output of Xor gate is shown above.

Buffer gate:

Let’s make it a buffer gate. Let any one of the inputs be 0 and other input be ‘x’.
So, for any value of ‘x’, the output ‘y’ is always same of ‘x’.

Thus acts as pass gate/buffer gate.

Inverter gate :

Now, let any one input to be ‘1’. Let other input be ‘X’ and output be ‘Y’.
So, now for any input ‘x’ the output ‘y’ will always be complement of ‘x’.

Thus it acts as inverter gate or complement gate.

Let’s understand with example.

Fix an input ‘a’ to be ‘1’.
Let , b=’1′.
Output = a(bar) b + b(bar) a = 1(bar) . 1 + 1(bar). 1 = ‘0’.
Let b=’0′
output = a(bar).b + b(bar).a = 1(bar).0 + 0(bar).1 = ‘1’.
So, if you notice, output is always complement of input ‘b’ when ‘a’ is always ‘1’.

Similarly you can try for buffer/pass gate also.

XOR gate can act an inverter as well as a buffer. If one fixes one input of the XOR gate to logic high it works as an inverter.

1 XOR 1=0

1 XOR 0= 1, So if one input of the XOR gate is fixed to logic 1 , the output is the logical inversion of the second input.

Now if one input of the XOR gate is fixed to logic zero it works as a buffer.

0 XOR 1=1

0 XOR 0=0, So if one input of inverter is fixed to logic 0 , the output is same as the second input.