Perceptron

A perceptron is a simple artificial neuron and one of the earliest building blocks of neural-network research. It multiplies input features by learned weights, adds a bias, and applies a threshold to choose between two classes. The perceptron learning rule updates weights when an example is classified incorrectly. A single perceptron can solve only linearly separable problems, meaning one straight decision boundary must divide the classes. Combining many units in multilayer networks with nonlinear activation functions overcomes that limitation and supports far more complex patterns. Although modern systems are much larger, the perceptron remains useful for explaining weights, biases, decision boundaries, and the foundations of supervised neural learning.