Patterns and Classification
Sorting items by a shared rule is classification. A computer can learn patterns from examples, and varied examples help us test whether its sorting works on new items.
What a learner can do afterwards
- Sort a set of items into categories and explain the rules they used
- Explain that computers learn patterns by looking at many examples
- Explain why varied examples and checks on new items matter when a computer learns a pattern
1 · Read
Sorting things into groups by a shared rule is called classification. You could put red blocks in one group and blue blocks in another. Your rule would be color. You could choose a different rule, such as shape.
Priya sorts 8 buttons. Each is round or square. There are 3 round buttons, so 8 minus 3 leaves 5 square buttons. Her rule is shape: round together, square together. A sorting rule should give each item a place. If she later finds a triangle button, she can add a triangle group.
A computer can learn a sorting pattern from examples. To train a new cat-and-dog picture sorter, people can give it pictures labeled 'cat' or 'dog'. It uses clues in those pictures to make guesses about new pictures. People write the software that lets it learn.
Cats can look different from each other. Varied pictures show more of those differences. A few pictures do not prove that a sorter works well on new cats. Check its guesses on pictures it did not train on. There is no fixed number of pictures that works for every sorter.
Classification groups things by a rule. Computers can learn sorting patterns from examples, but their guesses on new items need checking.
2 · Watch
Take it off screen
Where it sits
Where this leads
Jobs that lean on this skill. Follow one to see everything it is built on.
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.