An approach to discrete operator learning based on sparse high-dimensional approximation

Daniel Potts, Fabian Taubert

We present a dimension-incremental method for function approximation in bounded orthonormal product bases to learn the solutions of various differential equations. Therefore, we de…