Estimating per-unit-length resistance parameter in emerging Copper-Graphene hybrid interconnects via prior knowledge based accelerated neural networks

Abstract

In this paper, an artificial neural network (ANN) is developed to model how the geometrical parameters of hybrid copper-graphene interconnects affect the per-unit-length resistance values. The proposed ANN is intelligently trained using large amounts of data representing the prior knowledge about the interconnects, extracted from an analytical model and sparse amount of data extracted from a rigorous full-wave electromagnetic solver. In this way, the training of the ANN model is accelerated without significant loss in accuracy.