Neural networks originated from Donald Hebb’s 1949 theory that learning strengthens neural connections, “neurons that fire together, wire together.” This concept inspired computational models adjusting synaptic strengths. McCulloch and Pitts introduced the first artificial neuron in 1943, demonstrating logical operations. Later, Frank Rosenblatt’s 1958 perceptron introduced adjustable weights, enabling pattern recognition and bridging the fields of neuroscience, psychology, and computer science.
In 1969, Minsky and Papert’s critique exposed the flaws of single-layer perceptrons, particularly their inability to solve the XOR problem. With limited computational power, symbolic AI dominated, sidelining neural networks for over a decade. Their resurgence required new algorithms, such as backpropagation, showcasing how scientific paradigms evolve through breakthroughs.
