Perceptron is one of the simplest types of artificial neural networks, developed in the 1950s by Frank Rosenblatt (based on an invention of Warren McCulloch and Walter Pitts). It’s a single-layer model designed to perform binary classification — deciding whether an input belongs to one category or another based on weighted inputs.
A perceptron could take numerical features like “height” and “weight” to classify whether someone is “fit” or “not fit.”
Though basic, the perceptron laid the groundwork for more complex neural networks used in modern AI.