Neural Networks, this name appears to be something very much difficult to understand, something that can only be understandable by some super genius. Well, this is not that much difficult as it is supposed to be.
As we know human brains are incredibly genius in pattern recognition. We categorize the objects in the surroundings with little or no effort. The example normally given in pattern recognition is to understand the digits written by some human being.
Let’s say we have the digits written below and we have to recognize it with some computer
Thanks to https://towardsdatascience.com/ for copying the image. Let’s take an example of number 9 2nd last number from the right. We immediately judged number 9 without taking any time. Our train didn’t bother to take it as a small circle connected with a vertical line. Formulating rules for recognizing the handwritten numbers is extremely difficult.
Neural networks solve the problems as our brain do. They learn to recognize the patterns in the same way as we do but they do this by doing training on the labeled dataset.
First of all, let’s take the most basic artificial neurons; the Perceptron in the next part