Most of Al algorithms need to be trained before using. A regular training process which is repeated, evaluated, and refined may be named as 00000 Iterative Optimization Backpropagation O Gradient descent Loss function