Multilevel CNNs for Parametric PDEs based on Adaptive Finite Elements

TL;DR

This paper introduces a multilevel CNN architecture that leverages adaptive finite element meshes and error estimators to efficiently approximate parameter-to-solution maps of high-dimensional PDEs, matching state-of-the-art methods in accuracy and complexity.

Contribution

It presents a novel U-Net based CNN architecture that mimics multigrid algorithms and incorporates adaptive mesh refinement and error estimation for efficient PDE solutions.

Findings

Achieves comparable accuracy to low-rank tensor regression methods.

Reduces data complexity through adaptive mesh refinement.

Demonstrates practical effectiveness on UQ benchmark problems.

Abstract

A neural network architecture is presented that exploits the multilevel properties of high-dimensional parameter-dependent partial differential equations, enabling an efficient approximation of parameter-to-solution maps, rivaling best-in-class methods such as low-rank tensor regression in terms of accuracy and complexity. The neural network is trained with data on adaptively refined finite element meshes, thus reducing data complexity significantly. Error control is achieved by using a reliable finite element a posteriori error estimator, which is also provided as input to the neural network. The proposed U-Net architecture with CNN layers mimics a classical finite element multigrid algorithm. It can be shown that the CNN efficiently approximates all operations required by the solver, including the evaluation of the residual-based error estimator. In the CNN, a culling mask set-up according to the local corrections due to refinement on each mesh level reduces the overall complexity, allowing the network optimization with localized fine-scale finite element data. A complete convergence and complexity analysis is carried out for the adaptive multilevel scheme, which differs in several aspects from previous non-adaptive multilevel CNN. Moreover, numerical experiments with common benchmark problems from Uncertainty Quantification illustrate the practical performance of the architecture.