What is an Artificial Neuron?

Well, everyone has heard about Artificial Intelligence but we don’t have knowledge what happens in an Artificial mind (brain behind the intelligent Systems).

Neural Networks are the building blocks for making Intelligent Systems.

The inspiration of the Neural network came from the Biological neuron of Humans.

Let’s take time to understand the Architecture of our biological neuron which is very similar to Artificial Neuron

Biological Neuron

The above image shows a single biological neuron.

The input is received by the dendrites and is passed to main body of the neuron(Nucleus), the input signal is passed on towards axon which is basically a transmission line towards Axon terminals. The Axon terminals are connected to the dendrite of the next Neuron.

After understanding this biological neuron we will look into the structure of Artificial Neuron.

Artificial Neuron

Artificial Neuron

So this what the architecture of Artificial Neuron looks like.

Let’s gain the Understanding of this network.

The 3 arrows correspond to the 3 inputs coming into the network.

The values [0.7,0.6,1.4] are the weights assigned to the corresponding input.

So what are weights now?

Let’s take an example to understand this term.

Consider a university which grants student admission on 2 basis .One is their grades and second is the university test .

So this university has decided that grades will carry 70% of the overall weightage and university test will carry 30% for the admission process.

Coming back to our network.

Inputs get multiplied with their respective weights and their sum is taken.

Consider 3 inputs as x1,x2,x3.

Consider 3 weights be w1,w2,w3

Sum = x1w1 + x2w2+x3w3.

i.e. Sum=x1(0.7) + x2(0.6) + x3(1.4)

After summing we add bias to the sum obtained.

This Bias is just a constant number say 1 which is added for scaling purposes.

NewSum=x1(0.7) + x2(0.6) + x3(1.4) + bias

It’s not necessary to add bias but it is a good practice as it speeds up the process.

After adding bias, we reach at threshold step. If the newsum calculated is above the threshold value the neuron gets excited and it passes out the output.

If it doesn’t get excited it won’t pass on the output.

Stacking up many Artificial Neuron help us to create Intelligent Systems.

• Vidit Shah(Tuples)