A perspective on the next generation of invisibility cloaks—Intelligent cloaks

Invisibility has been a long-standing topic, owing to its crucial importance in both academia and industry. At the close of the last century, the advent of metamaterials unprecedentedly ignited the enthusiasm toward pursuing invisibility cloaks, but despite two decades of research, it has yet to mature into a practical method. In this Perspective, we first present a review of conventional and state-of-the-art invisibility methodologies, as well as their intrinsic challenges. To take a further step, we focus on the conjunction of two burgeoning disciplines—metamaterials and deep learning—to facilitate an intelligent self-adaptive invisibility cloak for dynamic surroundings. Finally, future directions are described, which include possible development directions and other practice-oriented metadevices driven by deep learning.