A covariance-based reduced-order framework for solving acoustic scattering problems (Q7288262)

We introduce an efficient covariance-based ROM framework for acoustic scattering problems by utilizing physics-aware reduced modeling and machine learning techniques to construct low-dimensional approximations of the contrast source density and its spatial covariance function. A non-intrusive GNN surrogate learns the complex mapping from scatterer geometry to the correlation kernel, significantly reducing computational cost while maintaining high accuracy in multiple-scattering scenarios.