Perceptron

Frank Rosenblatt, 1958

Invented by Frank Rosenblatt in 1958, the Perceptron is the foundation of neural networks. It takes weighted inputs, sums them, and passes the result through a step activation function. If the output doesn't match the target, weights are adjusted using a simple learning rule. The visualization shows inputs lighting up gold, weighted signals flowing to the summation node, activation threshold check, and weight updates after each training example.