Lesson 06 — What Is Machine Learning?

In the previous lesson, we saw that Machine Learning is a type of technology used to create AI systems.

But what does “machine learning” actually mean?

The basic idea is surprisingly simple.

What Is Machine Learning?

Machine Learning is a way of teaching computers to find patterns in data.

Instead of telling a computer exactly what to do in every possible situation, we give it examples and let it learn from them.

The computer uses these examples to build a model that can then make predictions or decisions about new data.

A Simple Example

Imagine that you want a computer to recognize cats.

One approach would be to manually program rules such as:

• Has two ears
• Has four legs
• Has whiskers
• Has fur

But this would quickly become complicated.

What happens if the cat is hidden behind something? What if the image is dark? What if the cat has unusual fur?

Machine Learning takes a different approach.

You give the system thousands of examples of cat and non-cat images.

The system looks for patterns in those examples.

After training, it can use what it learned to make predictions about new images.

The Three Basic Steps

A simple Machine Learning process can be divided into three main steps:

1. Give the system data.
2. Let it learn patterns from the data.
3. Use those patterns to make predictions.

That’s the basic idea behind Machine Learning.

Training Data

The examples used to teach a Machine Learning system are called training data.

For example, if you want to build a system that identifies spam emails, your training data could contain thousands of emails labeled as:

• Spam
• Not spam

The system looks at these examples and learns patterns associated with each category.

What Does the Computer Actually Learn?

The computer does not learn in exactly the same way a human does.

Instead, a Machine Learning algorithm adjusts mathematical parameters inside a model so that the model becomes better at making predictions.

You do not need to understand the mathematics to understand the main idea.

The model gradually adjusts itself to better match the patterns found in the training data.

Training vs Prediction

Machine Learning usually has two important phases.

Training

During training, the model receives examples and adjusts itself to improve its predictions.

Prediction

After training, the model can receive new data and use what it learned to make a prediction.

For example:

Training: The model studies thousands of emails.

Prediction: The model receives a new email and predicts whether it is spam.

Supervised Learning

One common type of Machine Learning is called supervised learning.

In supervised learning, the training data includes the correct answers.

For example:

• Email → Spam
• Email → Not spam
• Photo → Cat
• Photo → Dog

The model learns the relationship between the input and the correct result.

Unsupervised Learning

Another approach is called unsupervised learning.

Here, the data does not come with predefined answers.

The system tries to discover patterns or groups on its own.

For example, a company could give an algorithm information about thousands of customers.

The system might discover groups of customers with similar behaviors.

Reinforcement Learning

A third important approach is reinforcement learning.

Instead of learning from a collection of labeled examples, the system learns by interacting with an environment.

It receives rewards for good actions and negative feedback for bad ones.

Over time, it learns which actions tend to produce better results.

A simple example would be teaching an AI to play a game.

Winning could produce a reward, while losing could produce negative feedback.

Why Is Data So Important?

Machine Learning depends heavily on data.

If the training data is poor, incomplete, or biased, the resulting model can also perform poorly.

This is why the quality of the data is often just as important as the algorithm being used.

Can Machine Learning Make Mistakes?

Absolutely.

A Machine Learning model does not automatically understand the world.

It learns patterns from the data it receives.

If those patterns are incomplete or misleading, the model can make incorrect predictions.

This is one reason why Machine Learning systems need to be tested carefully.

Machine Learning in Everyday Life

You probably interact with Machine Learning systems every day without realizing it.

Examples include:

• Netflix recommending movies
• Spotify recommending music
• Email spam filters
• Search engines ranking results
• Fraud detection systems
• Online advertising
• Face recognition

Machine Learning is already deeply integrated into many digital services.

Machine Learning vs Traditional Programming

Traditional programming generally works like this:

Rules + Data → Output

With Machine Learning, the process is different:

Data + Expected Results → Model

Then:

Model + New Data → Prediction

This difference is one of the most important ideas to understand about Machine Learning.

Key Takeaways

• Machine Learning allows computers to learn patterns from data.
• Training data is used to teach a model.
• After training, the model can make predictions on new data.
• Supervised learning uses labeled examples.
• Unsupervised learning looks for patterns without predefined answers.
• Reinforcement learning learns through rewards and feedback.
• The quality of the data strongly affects the quality of the model.

What’s Next?

Now you understand the basic idea behind Machine Learning.

But how do these systems actually learn increasingly complex patterns?

In the next lesson, we will discover Deep Learning and see how neural networks changed the world of AI.