AI-assisted Infrared Photodetection, Thermal Emission and Spectrum Engineering for Sustainability

Abstract: This talk will report the recent research progress in our lab on AI-enabled optical spectrum optimization, engineering, detection and emission. We emphasize a lot on infrared regime, because infrared light contains huge thermal energy that may cause urban heating and infrared light on-chip detections usually needs cryogenic cooling. We establish a photonic paradigm assisted with AI algorithms, enabling to detect high-dimensional (intensity, polarization, and wavelength) infrared light, at a chip scale and at room temperature. Furthermore, we develop a passive cooling technology that is empowered by multiscale machine learning algorithms to co-design with materials/pattern/geometry, without any energy consumption. The prototype is further scaled up for cooling uniform, cooling paints for end-users in China and Singapore. Smart materials are further deployed to make the spectrum engineering tunable.