Variational Formulations of the Strong Formulation - Forward and Inverse Modeling using Isogeometric Analysis and Physics-Informed Networks

Kent-André Mardal, Jarle Sogn, Marius Zeinhofer

Abstract: The recently introduced Physics-Informed Neural Networks (PINNs) have popularized least squares formulations of both forward and inverse problems involving partial differential equations (PDEs) in strong form. We employ both Isogeometric Analysis and Physics-Informed Networks.