Simple definition
Machine learning is a way of building software that learns from examples instead of following only fixed, hand-written rules.
How it works
- Collect many examples, such as photos labelled "cat" or "not cat".
- A learning method adjusts itself until its guesses match the labels more often.
- The finished model is then used on new examples it has never seen.
Everyday example
In real life
A bank system learns what your normal spending looks like. When a payment looks very different, it flags it for a check.
Benefits and why this matters
Why this matters
Machine learning sits behind spam filters, recommendations, and fraud alerts. It helps to know that these systems learn from past examples.
- Handles patterns too complex to write as rules
- Improves when it sees more good examples
- Powers everyday tools like spam filters and recommendations
Risks and limitations
Things to watch out for
- Watch out: Poor or one-sided examples lead to poor or unfair results
- Watch out: It can fail in situations unlike the ones it learned from
- Watch out: It is often hard to explain why it gave a certain answer
Your beginner action
Try this
Open a streaming or shopping app and look at its recommendations. Ask yourself which of your past actions might have led to them.