“…For example, learning to classify known examples (i.e., supervised learning), or learning to recognize the characteristic structure of an object from input data with no additional information (i.e., unsupervised learning). 34,85,86 During learning, each synapse (i.e., the weight between layers) in the network is constantly predicted, strengthened, or weakened by the algorithm until an optimal setting has been reached. As a typical machine learning algorithm, ANN has achieved immense success in many fields, such as in image recognition, intelligent robots, automatic controls, prediction, estimation, and so on.…”