If you are completely new to these terms this post will give you a clear picture about them. Let us consider an example of a Toy store

Vicky owns a Toy store. He receives a bunch of toys all mixed up every week.

As it would be difficult to search for a particular toy if all are mixed up so, he asks his employees to segregate them manually.

Vicky observes that it is taking a lot of time to segregate them manually.

But he wants to scale fast

So he decided to use an AI powered automation robot to handle the task.  

AI is the effort to automate intellectual tasks normally performed by humans.

This is done based on a Rule-Based engine that has been hard coded by humans.

Here, The Robot scans the toy’s tag and identifies and then classifies them accordingly.

But what if the Toys come untagged

A Machine Learning –based algorithm is now proposed to solve the above problem.

To create a ML model features need to be extracted and defined for the model to train until it could recognize what each toy would look like

This is the way the ml algorithm works

Feature extraction->classification->output

Now vicky’s toy store is running successfully and more toys came in

But these were in large quantity and he had never seen these types of toys they were completely new to him.

How could he now extract this large data?

That is where Deep Learning comes into play

A DL-based algorithm is used to sort any toy by totally removing the need for defining features of each toy.

It doesn’t need to be provided with features to classify correctly ,It processes the provided images through neural networks (mimicing the human brain) to define specific features and classify them.

This is the way the DL algorithm works

Input->Feature Extraction +Classification->->output