In electromagnetic (EM) society, the forward/inverse and linear/nonlinear EM problems are usually solved via numerical algorithms such as the method of moments (MoM), finite-element method (FEM), finite-difference method (FDM), and finite volume method (FVM), etc. As an emerging discipline, scientific machine learning (SciML) has been applied to solve and model a variety of physical and engineering sciences. The success of SciML at various approximation tasks has motivated their potential in addressing EM problems. Thus, this session will focus on the advanced SciML based algorithms, theories and techniques in solving EM issues including but limited to the forward/inverse EM problems, linear/nonlinear EM problems, and its extensions to EM related multiphysics issues.