📝 SummarXiv

Abstract

Neural networks have been demonstrated to be able to provide fast surrogate models that can accelerate the inverse design of optical and electromagnetic devices. Nevertheless, such neural networks can be unreliable and normally require extreme amounts of data for training. Here we show that these limitations can be alleviated by constraining neural-network models using prior knowledge about the governing physics. We propose a universal physics-informed neural network for electromagnetic scattering based on the quasinormal mode expansion of the scattering matrix. Our neural-network model learns the resonant structure underlying the scattering spectrum and is guaranteed to obey energy conservation and causality. We demonstrate significantly improved data efficiency for photonic-crystal slabs and all-dielectric free-form metasurfaces. The method can be applied to a wide range of optical and electromagnetic devices owing to the generality of the quasinormal mode formalism.