בלוג

 Learn Python to A.I Programming – Lesson 5

The XOR Problem —  That problem stopped AI for 20 years.

1957–1969

• The perceptron, invented by Frank Rosenblatt (1957), was the first artificial neuron.

• It could learn simple logic like AND and OR, but not XOR.

• Marvin Minsky and Seymour Papert proved in 1969 (in Perceptrons: An Introduction to Computational Geometry) that a single-layer perceptron cannot solve XOR because it is not linearly separable.

• This proof killed early AI optimism. Funding for neural network research was cut.

• The 1970s–early 1980s became known as the “AI Winter” — progress froze for almost two decades.

The Solution — The Second Birth of AI

1986 onward

• Researchers like David Rumelhart, Geoffrey Hinton, and Ronald Williams discovered the backpropagation algorithm.

• Backpropagation allowed multi-layer perceptrons (MLPs) to train efficiently.

• With hidden layers and non-linear activations, MLPs could combine multiple linear cuts — solving XOR and all non-linear problems.

• This discovery unblocked AI completely.