Exact Representation of Piecewise Polynomial Finite Element Functions by Two-Hidden-Layer Multilayer Perceptrons

DOI10.48550/arXiv.2609.39856MaRDI QIDQ7361979zbMATH OpenarXivFDO

Authors Pengzhan Jin, Shengyang Xu, Jun Hu

Publication date 30 September 2026

Published in arXiv

Tldr This paper constructs exact representations of arbitrary polynomial degree piecewise linear finite element functions on simplicial meshes by two-hidden-layer multilayer perceptrons with a specific activation function. The construction is general for spatial dimensions, zero-skeleton and prescribed-skeleton cases, and continuous cases.

Full work available at URL https://doi.org/10.48550/arXiv.2609.39856

arXiv classification

math.NA

Cites work

• A family of conforming mixed finite elements for linear elasticity on triangular grids
• A mixed finite element method for 2-nd order elliptic problems
• A new family of mixed finite elements in \({\mathbb{R}}^ 3\)
• Approximation by superpositions of a sigmoidal function
• Better approximations of high dimensional smooth functions by deep neural networks with rectified power units
• De Rham compatible deep neural network FEM
• Deep ReLU networks and high-order finite element methods
• Exponential Expressivity of ReLU$^k$ Neural Networks on Gevrey Classes with Point Singularities
• Exponential expressivity of ReLU^k neural networks on Gevrey classes with point singularities.
• Finite element exterior calculus: From Hodge theory to numerical stability
• Finite element quasi-interpolation and best approximation
• Mixed finite elements in \(\mathbb{R}^3\)
• Neural Network Approximation
• Partitioning the projective plane and the dunce hat
• Polynomial interpolation in several variables
• PowerNet: efficient representations of polynomials and smooth functions by deep neural networks with rectified power units
• ReLU deep neural networks and linear finite elements
• ReLU Deep Neural Networks from the Hierarchical Basis Perspective
• Shallow ReLU neural networks and finite elements
• The Mathematical Theory of Finite Element Methods

This page was built for publication: Exact Representation of Piecewise Polynomial Finite Element Functions by Two-Hidden-Layer Multilayer Perceptrons

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7361979)