A neural operator framework for data-driven discovery of stability and receptivity in physical systems
A neural operator framework uses neural networks to automatically find stability and receptivity properties of nonlinear physical systems directly from trajectory data, bypassing the need for governing equations. The method trains a network to emulate system dynamics, extracts its local linear approximation (Jacobian), and then applies this data-driven operator to perform analyses like linear stability, Floquet analysis for periodic orbits, and resolvent analysis.