A novel orthogonal greedy algorithm (OGA) combined with shallow neural networks is proposed to efficiently solve fractional Laplace equations.
