Physics-embedded Fourier Neural Network for Partial Differential Equations
Not on the map. Why?
Abstract (arXiv)
We consider solving complex spatiotemporal dynamical systems governed by partial differential equations (PDEs) using frequency domain-based discrete learning approaches, such as Fourier neural operators. Despite their widespread use for approximating nonlinear PDEs, the majority of these methods neglect fundamental physical laws and lack interpretability. We address these shortcomings by introducing Physics-embedded Fourier Neural Networks (PeFNN) with flexible and explainable error control. PeFNN is designed to enforce momentum conservation and yields interpretable nonlinear expressions by utilizing unique multi-scale momentum-conserving Fourier (MC-Fourier) layers and an element-wise product operation. The MC-Fourier layer is by design translation- and rotation-invariant in the frequency domain, serving as a plug-and-play module that adheres to the laws of momentum conservation. PeFNN establishes a new state-of-the-art in solving widely employed spatiotemporal PDEs and generalizes well across input resolutions. Further, we demonstrate its outstanding performance for challenging real-world applications such as large-scale flood simulations.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 33 of its 58 references that are in the map. First posted on arXiv 15 Jul 2024.
Builds on
33 of 58 references.
References · 33
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
- 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,806
- 2016Group Equivariant Convolutional Networkscited by 1,154
- 2020Learning Mesh-Based Simulation with Graph Networkscited by 840
- 2021Physics-Informed Neural Operator for Learning Partial Differential Equationscited by 516
- 2021Machine learning accelerated computational fluid dynamicscited by 461
- 2019General E(2)-Equivariant Steerable CNNscited by 441
- 2021Equivariant message passing for the prediction of tensorial properties and molecular spectracited by 438
- 2022Fourier Neural Operator with Learned Deformations for PDEs on General Geometriescited by 416
- 2021Neural Operator: Learning Maps Between Function Spacescited by 413
- 2022PDEBENCH: An Extensive Benchmark for Scientific Machine Learningcited by 396
- 2022Message Passing Neural PDE Solverscited by 368
21 more
- 2017PDE-Net: Learning PDEs from Datacited by 304
- 2021Factorized Fourier Neural Operatorscited by 237
- 2018PDE-Net 2.0: Learning PDEs from Data with A Numeric-Symbolic Hybrid Deep Networkcited by 208
- 2023Convolutional Neural Operators for robust and accurate learning of PDEscited by 207
- 2017Beyond Finite Layer Neural Networks: Bridging Deep Architectures and Numerical Differential Equationscited by 203
- 2020PhyGeoNet: Physics-Informed Geometry-Adaptive Convolutional Neural Networks for Solving Parameterized Steady-State PDEs on Irregular Domaincited by 173
- 2020Incorporating Symmetry into Deep Dynamics Models for Improved Generalizationcited by 123
- 2022Physics-Informed Deep Neural Operator Networkscited by 110
- 2020Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamicscited by 93
- 2022AirfRANS: High Fidelity Computational Fluid Dynamics Dataset for Approximating Reynolds-Averaged Navier-Stokes Solutionscited by 77
- 2018Deep Dynamical Modeling and Control of Unsteady Fluid Flowscited by 67
- 2021Encoding physics to learn reaction-diffusion processescited by 61
- 2023Group Equivariant Fourier Neural Operators for Partial Differential Equationscited by 42
- 2019Physical Symmetries Embedded in Neural Networkscited by 40
- 2022Physics-Embedded Neural Networks: Graph Neural PDE Solvers with Mixed Boundary Conditionscited by 33
- 2022Guaranteed Conservation of Momentum for Learning Particle-based Fluid Dynamicscited by 30
- 2022INO: Invariant Neural Operators for Learning Complex Physical Systems with Momentum Conservationcited by 28
- 2022Exact conservation laws for neural network integrators of dynamical systemscited by 15
- 2023Physics-aware Machine Learning Revolutionizes Scientific Paradigm for Machine Learning and Process-based Hydrologycited by 11
- 2023Equivariant Neural Operator Learning with Graphon Convolutioncited by 10
- 2024Large-scale flood modeling and forecasting with FloodCastcited by 5
Built on by
Citing papers per year: 2026: 1.
1 paper.