Aimed at senior undergraduate or first-year graduate courses in neural networks and neurocomputing, this work presents neural network theory for diverse applications in a unified way, where the structures of artificial neural networks are characterized by distinguished classes of graphs.