Subwavelength metamaterial devices with optimization and machine learning

Subwavelength metamaterials allow to synthesize tailored optical properties which enabled the demonstration of photonic devices with unprecedented performance and scale of integration. Yet, the development of metamaterial-based devices often involves a large number of interrelated parameters and figures of merit whose manual design can be impractical or lead to suboptimal solutions. In this invited talk, we will discuss the potentiality offered by multi-objective optimization and machine learning for the design of high-performance photonic devices based on metamaterials. We will present both integrated devices for on-chip photonic systems as well as recent advances in the development of devices for free-space applications and optical beam control.