AI‐Driven Approaches for Solving Electromagnetic Inverse Problems

This chapter provides an overview of artificial intelligence-driven methods for solving electromagnetic (EM) inverse problems (IPs) with high reliability, robustness, and computational efficiency. Several methodologies are detailed and discussed, including the recent developments within the so-called (i) three-step learning-by-examples, (ii) system-by-design, and (iii) deep learning frameworks. Afterward, a survey on their customization to several EM–IPs, such as those arising in microwave imaging of free-space and buried targets, biomedical imaging, nondestructive testing and evaluation, as well as the detection, localization, and tracking of non-cooperative targets in wireless networks, is given.