Global Feature Extractor As mentioned in the paper, the global feature extractor GFE()is composed of 3 modules: GFE(Mn) = GAP(Conv(Sample(Mn))). The sampling module Sample()is implemented by the built-in interpolation interface in Firedrake[Rathgeber et al., 2016], with sampling density 32x32. The convolutional block contains 4 convolutional layers. The SELU [Klambauer et al., 2017] activation function is used to increase the representation capability of the model. The output tensor from the convolutional block is then fed into a Global Average Pooling (GAP) layer to get a mesh resolution invariant global feature embedding En.

Classically, to solve differential equation problems, it is necessary to specify sufficient initial and/or boundary conditions so as to allow the existence of a unique solution. Well-posedness of differential equation problems thus involves studying the existence and uniqueness of solutions, and their dependence to such pre-specified conditions. However, in part due to mathematical necessity, these conditions are usually specified "to arbitrary precision" only on (appropriate portions of) the boundary of the space-time domain. This does not mirror how data acquisition is performed in realistic situations, where one may observe entire "patches" of solution data at arbitrary space-time locations; alternatively one might have access to more than one solutions stemming from the same differential operator. In our short work, we demonstrate how standard tools from machine and manifold learning can be used to infer, in a data driven manner, certain well-posedness features of differential equation problems, for initial/boundary condition combinations under which rigorous existence/uniqueness theorems are not known. Our study naturally combines a data assimilation perspective with an operator-learning one.