Gradient Descent

ML

An optimization algorithm used to train machine learning models by iteratively updating model parameters in the direction that reduces the loss function. The gradient indicates the steepest direction of increase, and the algorithm moves in the opposite direction.