AI‐Driven Approaches for Solving Electromagnetic Inverse Problems

Abstract

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.