Prepared the Automatic Differentiation documentation covering computational graphs, forward and reverse-mode differentiation, chain-rule derivatives, gradient accumulation, broadcasting, neural network backpropagation, numerical gradient verification, performance, and testing.