{"id":"arxiv/2608.27256","url":"https://arxiv.org/abs/2608.27256","tags":["math.NA","cs.LG"],"score":0,"title":"Enforcing Dirichlet Boundary Conditions in Operator Learning","domain":"math.NA","source":"arxiv","authors":["Andrew M. Stuart","Margaret Trautner"],"reviews":0,"subject":"math.NA","summary":"Operator learning in scientific machine learning is concerned with approximation of maps between infinite-dimensional function spaces; such maps frequently arise as the solution operators of partial differential equations (PDEs). Neural operators have demonstrated broad empirical success at approximating such maps from data. However, most existing neural operator architectures enforce boundary conditions indirectly through training from data even though the boundary condition is often known exactly. Furthermore, existing modifications and approaches that do enforce boundary conditions…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-08-27","reactions":{}}