You might've heard about machine learning and artificial intelligence when you start learning computer science – so did I.

I'm pretty sure there are some things that you don't really understand, like how can a machine be intelligent, and how do actually machine learns?

Well, in this article, originally posted as a Twitter thread, I'm trying my best explaining about how does it all work, specifically, how do machine learns.

So let's get into it –

First we look at how human learns, oftentimes, we learn from our past experiences and practice. Machine learns kinda the same way, but with data. A lot of data.

There are 2 types of learning:

• Supervised Learning
• Unsupervised Learning

Supervised Learning

For supervised learning, the machine is given some labelled data. The goal of supervised learning is to make predictions accurately.

For example, supervised learning for a human.

• A kid was told by his teacher that a monkey looks like this 🐒
• And a horse looks like this 🐎

After several different photos of monkeys and horses has been shown, the kid will start recognise the pattern.

Say monkeys has a pair of legs and limbs, while horses have 4 legs.

Then, if the teacher shows a set of other different photos of monkeys and horses, even it has never been shown before,

Chances are, the kid will guess accurately. We can say now that the kid has "learned".

Same goes to a machine, but machine uses what's called Artificial Neural Network to learn.

I have written an article about Artificial Neural Network here, feel free to read it.

Unsupervised Learning

Another type of learning is called unsupervised learning, and as you may guess, the machine is given a set of unlabeled data.

The goal of this is to classify stuff – to see a pattern.

Again, to explain this, I'm going to use human as an example.

This example may be confusing to imagine but try to assume that we don't already know what a bird and a cat looks like.

If we are given a set of photos of some unidentified species, we will start to recognise some patterns.

For example the photos consists of 2 different unidentified species.

We may classify them by their colours, or number of legs, or number of the eyes.

Say species A is a type of bird and species B is a type of cat.

Even though we have never seen them before, we are able to classify them or group them based on their features.

From there, we learned that the set of photos given has 2 different species – a bird species and a cat species.

You may be thinking – "What if one of the photos classified as 'cats' is actually a fish?"

Well, going back to the goal of unsupervised learning. It is not about the accuracy, it's about the classification.

The real answer might not be accurate but we recognize a pattern. It is in fact that these 2 species are different, based on their features, the pattern.

Conclusion

Both of these type of learning is done by machines with a method called Artificial Neural Network, ANN.

Again, you can read about it here.

But in this post, I hope you understand the difference between supervised and unsupervised learning.

Next time we're gonna look at how machine uses ANN to learn just like humans.

I hope you're excited, hahaha.