Learning Incoherent Light Emission Steering From Metasurfaces Using Generative Models
Date:
Abstract:
This paper introduces an active-learning framework that drives a generative model to discover optimal pump patterns for steering incoherent photoluminescence from reconfigurable semiconductor metasurfaces. We achieve an order-of-magnitude improvement in steering efficiency compared with human-designed patterns, demonstrating a powerful combination of machine learning and nanophotonics for directional incoherent emission control.