Machine learning is the way most modern AI learns. Instead of being told every rule, the computer learns from examples and gets smarter over time.
It works in three steps: you give it lots of examples, it looks for patterns, and then it uses those patterns to make a guess.
Imagine teaching a computer to tell cats from dogs. You show it many labeled pictures. It learns the patterns. Then it sees a new picture and guesses, "That's a cat!"
Machine learning helps with all kinds of things: sorting photos, spotting patterns in data, recommending shows, and making predictions like the weather.
It can go wrong if the examples are bad, unfair, or confusing. Wrong labels teach the wrong thing, and one-sided examples can be unfair.
So good examples really matter. Smart machines learn from good examples, the same way curious humans do.
What to remember
- Machine learning means learning from examples.
- The steps: examples in, find patterns, make a guess.
- It powers photo sorting, recommendations, and predictions.
- Bad or unfair examples lead to bad guesses.
Words to know
- Machine learning
- Teaching computers by example instead of fixed rules.
- Example
- One piece of data the computer learns from.
- Label
- The correct answer attached to an example.
- Bias
- Unfairness that creeps in from one-sided examples.
For grown-ups
Machine learning fits models to data so they generalize to new inputs. Supervised learning uses labeled examples; the quality, balance, and representativeness of that data largely determine performance and fairness. 'Garbage in, garbage out' is literal here, which is why data curation and bias evaluation matter.
Want the full story? These go deeper:
