🐱 Machine learning: learning from examples

Unit 1 · How machines learn · 10 min

To teach a computer cats vs dogs, give it thousands of photos labelled “cat/dog” (training data). The model guesses, checks, adjusts, again and again, until it is accurate. Then test it on photos it has never seen. More varied data means a more reliable model; biased data means a biased model.

Words in this lesson

• model — model
• training data — training data
• bias — bias

Project: Be a human model: collect 10 different leaves, invent a rule that splits them into two groups, then have a friend sort new leaves with your rule and check how well it works.