Geometry-aware LegONet for PDE Learning on Arbitrary Domains

DOI10.48550/arXiv.2607.23069MaRDI QIDQ7284438arXivFDO

Authors Jiahao Zhang, Yueqi Wang, Guang Lin

Publication date 25 July 2026

Published in arXiv

Tldr We introduce a learned PDE solver that separates physical-law discovery from geometry and allows reuse across different geometries by pretraining modular blocks and assembling them in target geometries via boundary constraints; this achieves predictive governing laws and nearly zero residual errors on unseen domains.

Full work available at URL https://doi.org/10.48550/arXiv.2607.23069

arXiv classification

math.NA

Cites work

• DGM: a deep learning algorithm for solving partial differential equations
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks
• Fourier Neural Operator for Parametric Partial Differential Equations
• Fourier Neural Operator with Learned Deformations for PDEs on General Geometries
• Geometric Numerical Integration
• Geometry-Informed Neural Operator for Large-Scale 3D PDEs
• GNOT: A General Neural Operator Transformer for Operator Learning
• LegONet: Plug-and-Play Structure-Preserving Neural Operator Blocks for Compositional PDE Learning
• Matrix Computations
• On the Construction and Comparison of Difference Schemes
• Physics-informed geometry-aware neural operator
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Regression Shrinkage and Selection Via the Lasso
• Spectral Methods in MATLAB
• The Mathematical Theory of Finite Element Methods
• Weak SINDy For Partial Differential Equations

This page was built for publication: Geometry-aware LegONet for PDE Learning on Arbitrary Domains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284438)