Continuous Representation Learning for Spatiotemporal Super-Resolution of Parametric PDE Solutions

The proposed Super Resolution Operator Network (SROpNet) framework learns continuous representations of solutions to parametric differential equations from low-resolution numerical approximations, enabling spatiotemporal super-resolution without constraints on sensor and prediction locations.