Introduction to Emerging Technologies
Primary 6 · Science and Elementary Technology
Introduction to Emerging Technologies
Lesson 4 of 5
25 min
Let's begin
Habimana sells fruit at Kimironko market. His little sister is four years old and is learning to tell mangoes from avocados. Habimana does not give her a list of rules. He simply holds up fruit after fruit and says, "mango", "avocado", "mango". After seeing many examples, she can sort new fruit all by herself.
Computers can learn in a very similar way. This is called machine learning.
Machine learning is a kind of artificial intelligence in which a computer learns from examples instead of being given every rule by a programmer. The computer scientist Arthur Samuel described it in 1959 as giving computers "the ability to learn without being explicitly programmed". From the examples, the computer builds a model, and the model makes predictions (good guesses) about new things.
Imagine writing a rule: "If the fruit is green and its skin is bumpy, it is an avocado." But some mangoes are green, some avocados are dark purple and smooth, and fruit looks different in bright sun and in shade. You would need hundreds of rules. It is easier to show the computer many examples and let it find the patterns itself.
| What you need | What it means | Mango and avocado example |
|---|---|---|
| Data | Many examples | Lots of photos of mangoes and avocados |
| Labels | The correct answer for each example, given by people | Each photo is marked "mango" or "avocado" |
| A learning program (algorithm) | Instructions that search the examples for patterns | It compares shapes, colours and skin |
| A computer | Runs the learning program | A school laptop connected to a learning website |
| A model | What the computer builds after learning; it makes predictions | "This new photo is 90% likely to be a mango." |
| Test data | New examples, never used for learning, to check the model | Fresh photos taken by another group |
A feature is a clue that helps tell things apart.
| Feature | Mango | Avocado |
|---|---|---|
| Shape | Oval or kidney-shaped, often with a small point | Pear-shaped or round |
| Skin | Smooth | Often bumpy or leathery |
| Colour | Green, yellow, orange or red | Green to dark purple |
Notice that colour alone is not enough: both can be green! That is why the model needs many examples.
Tip
Good examples in, good guesses out. A model is only as good as the examples it learns from. If the examples are few, one-sided or wrongly labelled, the guesses will be poor too.
Answers
What you learned today
Sources and credits
The exercises (Try it yourself and the practice questions) are written by Ganzaa.
This lesson adapts CC BY-SA material and is shared under CC BY-SA 4.0.
Quick check
Match each machine learning word to its meaning.
Tap one on the left, then its partner on the right.
