Now let me ask you a thing , In which field we are better than the computers . Not that love and anger things 🤣 . We humans are good at recognising the objects or some image whereas the computer is best at doing the algorithm work .
So now if I ask a machine that which among these is 7 or 9 or 1 or 3
then answer cant be that precise because we dont have any type of process for that ..
Here the neural network come into the play
Neural Networks create their own rules on the basis of dataset . Like in this case if we give this dataset to the machine and ask after that in our own handwritting than the computer would be able to answer it . because we have provided him/her about different types of numbers . Now from where we would get these datasets ?
Modified National Institute of Standards and Technology dataset
The MNIST dataset is an acronym that stands for the Modified National Institute of Standards and Technology dataset. It is a dataset of 60,000 small square 28×28 pixel grayscale images of handwritten single digits between 0 and 9.
It have different fields datasets too. Like for fashion , cars , animals and many more !!
So as when studying for human body we begin it with cells , for neural network we have perceptron here
These are the building blocks of Machine Learning. A neural network unit that does certain computations to detect features or business intelligence in the input data
This is how a perceptron works ! Firstly we gave multiple inputs which would be binary ( 0 or 1 / True or False ) and the weight of them which are represented as " w " which shows the importance of that in the result !
y -> a function that would do everything to find out this result !!
We have an output .. This was from the surface . How it looks and how it goes .
Now here is something we discussed above but a new word came here , threshold ...... What's that ? It's the requirement of the user from the output , like if we want a black shirt and a full shirt , then for black it would be 1 and for that too it would be1 Simply the required weight in output is threshold !
So for each input we do input * weight  ...... till mth input and add that up if that is > threshold then output is 1 otherwise 0 !