A comparative study of local and global preconditioners for finite element analysis

A Finite Element based Deep Learning solver for parametric PDEs

“…In practice, the inverse of A i is non-computable and thus the error is never known. Hence, as in iterative methods for solving a system of linear equations, we select a preconditioner to mimic the inverse operator and thus decrease the condition number of the system [39]. Thus, we define the loss as…”