NanoPhotoNet-MPM: Physics-Informed Neural Surrogate for Inverse SPDC Design
A physics-informed neural surrogate framework combining an EigenmodeDeepONet transverse eigensolver, a deterministic physics conversion layer, and a CWE-PINN longitudinal propagator for inverse design of modal phase-matched biphoton quantum light sources in anisotropic monoclinic NbOCl2 waveguides.
Methods: Physics-informed neural networks (PINNs), DeepOperator Networks (DeepONet), coupled-wave equations, finite-difference eigenmode (FDE) solvers, genetic algorithm optimization, and joint spectral amplitude (JSA) analysis.