Machine Learning

AI is one of the most widely used words in recent times. So, what exactly is AI?

Artificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines.- Wikipedia

In order to get intelligence or be intelligent, one must learn. “Machines, when trying to replicate or simulate such learning is what machine learning in AI is”

Most people who’ve heard about Machine Learning would have come across this definition.

A Field of study that gives computers the ability to learn without being explicitly programmed. — Arthur Samuel.

Evidently, Arthur Samuel coined the term “Machine Learning”.

Now a question arises — what is explicitly programmed?

Formulating the rules or the codes manually by a programmer, where every part of that system is instructed by the programmer.

So, How is Machine learning different from this?

In Machine learning, They will identify and match the patterns with the help of algorithms from the given data.

Based on how they identify and match patterns, We can classify them into three categories.

1. Supervised Machine Learning
2. Unsupervised Machine Learning
3. Reinforcement Machine Learning

Supervised Machine Learning:

Here, Machines learn from previous Inputs (Independent Variable) and Outputs (Dependent Variable) based on the data we provide.

Let’s assume we need to automate the process of separating the emails we receive in the future. (Spams and Non-spams)

The previous emails we received are Inputs (Independent Variables) and the labels we have given to them like Spam and Non-spam are outputs. (Dependent Variables).

Based on this given information, we will train a Supervised Machine Learning algorithm to predict upcoming emails as Spam or Non-spam.

That means we are educating the machine on what is spam email and what isn’t.

For example:

For a 2-year old we can teach fruit names based on their appearance factors like their shape, color and etc. Based on what we taught the 2-year old can classify the fruits on their own.

It is a classical Example of Garbage in, Garbage out.

We get what we feed to the machine. High-quality data will provide high-quality output.

Unsupervised Machine Learning:

Here, the machine will find trends and patterns between Inputs (Independent Variables) without reference to known labels or outcomes.

For our mail automation process, if we have the emails and not their labels, we can train an Unsupervised Machine learning algorithm for these emails and predict labels accordingly. The machine identifies the patterns between the emails and segments them based on their patterns/trends. Then we can use that segmented data for further predictions.

For Example:

Let’s suppose you are watching a football match for the first time and without prior information about each player you can segment them based on the T-shirts they wear.

How is this learning different from Supervised Learning? Since you didn’t use any past/prior knowledge about players and classified them “on-the-go”.

Reinforcement Learning:

It is another type of Machine Learning, which focuses mainly on making models that learn from their mistakes.

For example:

While playing a game, first we make mistakes then we learn and adapt. In fact, that’s how we actually learn everything. Reinforcement Learning uses the same technique to train its brain (model) with its agent, Rewards and Action.

I hope you were able to understand what Machine Learning is and the different types it encompasses. There’s a lot more to write about this topic, so stay tuned for my next blog!

Have Fun :)