Machine learning
What is Machine learning?
Software that learns patterns from examples instead of following rules a person wrote by hand. You feed it data, it works out the patterns, and then it makes predictions about new data it hasn't seen. Most of what people call "AI" today is machine learning underneath.
Why it matters
Traditional software does exactly what a programmer wrote, rule by rule. Machine learning flips that. You show the system many examples and it works out the patterns itself, then applies them to data it has never seen. This matters because a huge number of real problems are too messy to write rules for. Nobody can list every way a fraudulent transaction looks, but a model trained on millions of past transactions can spot the pattern. Most of what people now call “AI” is machine learning underneath, so understanding this one idea explains most of the field.
In practice
A bank trains a model on years of transactions labelled fraud or not, and it learns to flag suspicious ones in real time. A retailer predicts next month’s demand from past sales. A postal service reads handwritten postcodes. In every case nobody wrote the rules by hand. They supplied the examples, the model found the pattern, and its accuracy rises or falls with the quality of that data.
