Partial differential equations (PDE) must be solved repeatedly for different values of some parameters to address several science and engineering problems, including in molecular dynamics and micro-mechanics. Existing approaches to solving PDE like finite element methods and finite difference methods are typically slow and inefficient. Li et al. present the Fourier neural operator, a neural operator architecture defined directly in the Fourier space. The Fourier neural operator, which outperforms existing deep-learning methods, can learn an entire family of PDEs from scratch at any resolution far faster than traditional solvers.