Recent Advances in Deep Neural Network Technique for High-Dimensional Microwave Modeling
Jing Jin, Feng Feng, Wei Zhang, Jianan Zhang, Zhihao Zhao, Qi‐Jun Zhang
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.
Topics & Concepts
Artificial neural networkDeep learningComputer scienceSigmoid functionArtificial intelligenceMicrowave imagingDeep neural networksTime delay neural networkMicrowaveTelecommunicationsIndoor and Outdoor Localization TechnologiesSoil Moisture and Remote SensingElectromagnetic Simulation and Numerical Methods