Uncertainty analysis on FDTD computation with artificial neural network

The artificial neural network (ANN) has appeared as a potential alternative for uncertainty quantification (UQ) in the finite difference time domain (FDTD) computation. It is applied to build a surrogate model for the compute-intensive FDTD simulation and to bypass the numerous simulations required for UQ. However, when the surrogate model utilizes the ANN, a considerable number of data is generally required for high accuracy and generating such large quantities of data becomes computationally prohibitive. To address this drawback, a number of adaptations for ANN are proposed which additionally improves the accuracy of the ANN in UQ for the FDTD computation while maintaining a low computational cost. The proposed algorithm is tested for application in bioelectromagnetics and considerable speed-up, as well as improved accuracy of UQ, is observed compared to traditional methods such as the non-intrusive polynomial chaos method.