Evolution of Back Propagation: Who Invented It?

Explore the historical development of backpropagation in this 27-minute educational video that traces the algorithm's origins and evolution. Discover why perceptron convergence learning became obsolete, why analytical methods proved impractical, and how numerical methods like gradient descent offered viable alternatives. Learn about gradient vectors and the direction of steepest ascent, efficient gradient computation techniques, and the meaning behind the "backward" in backpropagation. The video examines the contributions of key figures including Cauchy (who invented gradient descent in 1847), Rumelhart, Williams, Hinton, Paul Werbos, and others who developed this fundamental neural network training algorithm. Complete with extensive resources including original papers, tutorials, and supplementary materials, this comprehensive exploration concludes with a quiz and summary to reinforce understanding of backpropagation's historical development and mathematical foundations.