Gradient Descent
An optimization algorithm that iteratively adjusts model parameters to minimize a loss function. It works by computing the gradient (direction of steepest increase) and moving parameters in the opposite direction. Variants include SGD and Adam.