wunder beta

🤖 How Machine Learning Works

Computers can learn — not by being told every rule, but by looking at lots of examples, finding the pattern, and then making a guess. Learn the three big steps behind machine learning, and see why mac

lessons

~15 min

to learn

🔬 Science

subject

Adults

level

Start the course →

What you’ll learn

Learning From ExamplesUnderstand that machines can learn from examples, much like people do.People aren't handed a rulebook for recognizing a dog — they learn from many examples. A learning machine works the same way: it studies example data instead of being told every rule, then gets better at guessing the right answer.
Step 1: Show Lots of ExamplesLearn that training means giving a machine many labeled examples.The first step is to show the machine many examples, each with the correct answer attached — a label. More examples, and more varied ones, lead to better learning; a couple of examples is never enough.
Step 2: Find the PatternSee that learning means spotting shared clues in the examples.After seeing labeled examples, the machine hunts for patterns — the clues that go with each answer, like pointy ears for cats or floppy ears for dogs. Finding these shared features is the core of how it learns.
Step 3: Make a PredictionConnect the three steps and see machine learning in daily life.Once it has learned the pattern, the machine can make predictions about brand-new things. Show examples, find the pattern, make a prediction — those three steps power video suggestions, voice helpers, spam filters, and photo apps.
When Machines Get It WrongRecognize that machines copy their examples and can make mistakes.Machines don't truly understand; they copy the patterns in their examples. Missing or one-sided examples lead to mistakes, like an only-brown-dogs machine failing on a white dog. Giving more varied examples and feedback helps them improve.

Questions this course answers

How does a machine that learns figure things out?

Like a child learning what a dog is, a learning machine studies many examples and works out the pattern.

Learning from examples is most like:

You learned 'dog' from many examples — machines learn the same way, from lots of example data.

In machine learning, what is a 'label'?

A label is the correct answer given with each training example, like 'this is a cat.'

To train a machine well, you should give it:

The more examples — and the more varied they are — the better the machine can learn.

When a machine 'finds the pattern,' it is:

Finding the pattern means noticing the shared clues that point to each answer.

What are the three steps of machine learning, in order?

First show examples, then find the pattern, then use it to make a prediction on something new.

Grounded in trusted sources

MIT — 'How To Grow a Robot' / MIT Media Lab explainers on machine learning for kids
Google — 'Teachable Machine' project and explainers (teachablemachine.withgoogle.com)
Wikipedia — 'Machine learning' and 'MNIST database' (en.wikipedia.org)

Every Wunder lesson is built from real, reputable sources — never invented.

Related Science courses

🐸How Rainforest Animals Survive 🌿How Rainforest Plants Survive 👉How Recommendations Work 🔊How Voice Assistants Work 🤖Humans vs. Machines 🌳Indigenous Rainforest Knowledge 🌴Inside a Rainforest 🤖Making Rules for AI

Wunder is a personalized learn-anything platform — tell it any topic and it builds a beautiful, fact-checked course in minutes, with narration, a knowledge check, and a college-style University track.

Browse more Science courses · All topics · Home