Introduction
The perceptron, invented by Frank Rosenblatt in 1957, is the grandfather of all neural networks. It's one of the simplest models you can imagine — and studying it shows, in miniature, how neural networks make predictions and how they learn from mistakes.
Picture a strict bouncer at a club deciding whether to let someone in. They consider a few things — age, dress code, guest list — give each a different importance, add them up, and let the person in only if the total passes a certain bar. That's a perceptron.