McCulloch-Pitts Neuron

McCulloch-Pitts Neuron it is known as Mankinds first Algorithm for deep learning. It is based on normal mathematical functions/operations to train the program. It is basically based on neurons of Human brains.

Dendrite: Receives signals from other neurons
Soma: Processes the information
Axon: Transmits the output of this neuron
Synapse: Point of connection to other neurons

McCulloch-Pitts Neuron

It uses binary activation function which means the inputs should be only in 2 different numbers 0 or 1 and the algorithm will give results in the form of 0 and 1 only.

• The firing state is binary (1= fire, 0=not fire).
• Weights in the network remain the same throughout the program.
• It was able to solve Linear separable problems.
• There is no bias. This algorithm is known as the unbiased deep learning algorithm.
• It was able to solve Logical Gates AND, OR, NOT.

AND GATE:

if(sum)>threshold:
output=1
else:
output=0

Steps to solve manually:

1. Assume weight = 0(discard because weight cannot be 0 if we multiply it with any given input it will result in 0 itself).
2. Assume weight = 1.
3. Using the formula we will calculate threshold value using the input x1,x2 from the input of AND gate.
4. x1w+x2w= 1x1+0x1
5. From step 4 we can get the value 2 which is value for the threshold.

NAND GATE 

Let’s solve NAND Gate using the same steps above:

Let’s assume w=-1 so when x1=x2=0.

yin=x1w1+x2w


(0x-1 + 0x-1=0)

when

x1=1

x1=1 =>1x-1+0x-1= -1

x2=0

x1=0 0x-1+1x-1=-1

x2=1

x1=11x-1+1x-1=-2

x2=1

Result:

 0   1    1

-1   1    1

-1    1   1

-2    0   0

the algo:

if yin>=-1
then y=1
else:
y=0

Your task is to solve OR, NOT Gate using the steps above.