PINN-FEM enforces Dirichlet boundary conditions in PINNs by blending neural network fields with finite element shape functions in a boundary layer, but the 2D extension is ambiguous and the experimental comparisons are confounded.
Machine learning and big scientific data
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
PINN-FEM: A Hybrid Approach for Enforcing Dirichlet Boundary Conditions in Physics-Informed Neural Networks
PINN-FEM enforces Dirichlet boundary conditions in PINNs by blending neural network fields with finite element shape functions in a boundary layer, but the 2D extension is ambiguous and the experimental comparisons are confounded.