Feb, 2024
关于参数到可观测映射的操作学习视角
An operator learning perspective on parameter-to-observable maps
TL;DRFourier Neural Mappings (FNMs) framework introduces computationally efficient surrogates for parametrized physical models using the operator learning perspective, accommodating finite-dimensional inputs and outputs and demonstrating benefits for approximating nonlinear parameter-to-observable (PtO) maps.