The Levenberg–Marquardt algorithm typically converges rapidly to the local minimum, and appears to be the fastest method for training moderate-sized MLPs. The regularization refers to a process of introducing additional information during training, in the form of a penalty for model complexity.