Artificial intelligence, as a key tool for data-driven science advancements, has significantly accelerated computational discovery. A salient example is computational electromagnetics, whereby increasingly sophisticated machine learning algorithms and artificial neural networks are rapidly making a highly disruptive impact in this area of computational science. In particular, by interweaving AI-driven algorithms with numerical methods widely used in computational electromagnetics it has become possible to tackle highly nontrivial inverse problems pertaining to the design of functional photonic structures, including intelligent metamaterials, active photonic nanodevices, and reconfigurable metasurfaces. This session is devoted to recent advances in computational electromagnetism, whereby concepts and techniques pertaining to artificial intelligence, machine learning, and deep learning are seamlessly integrated with computational methods central to electromagnetics modelling leading to new improved or vastly superior computational methods.