How machines learn
Supervised learning is machine learning where the model trains on examples that come with the correct answers, so it can learn to predict answers for new data.
Imagine flashcards with a picture on the front and the answer on the back. Supervised learning is training with a huge stack of those cards. Each example has a label: this photo is a cat, this email is spam, this house sold for this price. The model guesses, checks the label, and adjusts.
The word 'supervised' refers to the labels acting like a teacher. Most practical AI in business is supervised: fraud detection, medical image screening, price prediction. The catch is that someone has to create all those labels, which is slow and expensive at scale.
Supervised learning splits into two big jobs. Classification predicts a category, such as spam or not spam. Regression predicts a number, such as a price. Both need labeled data, and both are judged by how well they do on examples they never saw during training.
A bank's fraud model learns from millions of past card transactions that staff had already marked as genuine or fraudulent, then applies that to new purchases.