Recent Advances in Deep Neural Network Technique for High-Dimensional Microwave Modeling

Abstract

This paper provides an overview of recent advances in deep neural network technique for high-dimensional microwave modeling. The hybrid deep neural network that employs both the sigmoid function and the smooth rectified linear unit (ReLU) as activation functions is used for microwave modeling in order to address the challenges due to high-dimensional inputs. A three-stage deep learning algorithm is used to train the deep neural network model. It can overcome the vanishing gradient problem for training the deep neural network. The deep neural network technique can solve microwave modeling problems in higher dimension than the shallow neural network method.