Reference · Glossary
Machine Learning
A branch of AI where programs **improve from examples** instead of following only hand-written rules.
When to use
When something gets better with more data: spam detection, product recommendations, fraud alerts, speech-to-text.
When not to
When a simple if/then rule is enough ("if password wrong three times, lock account") — that's not ML.
Example
Show 10,000 photos labeled "cat" or "not cat." The model learns patterns so it can guess on new photos it has never seen.