Learning the Willmore flow with a lightweight neural operator

DOI10.48550/arXiv.2610.03534MaRDI QIDQ7374850zbMATH OpenarXivFDO

Authors Elie Bretin, Roland Denis, Tokuhiro Eto, Simon Masnou

Publication date 2 October 2026

Published in arXiv

Tldr We extend a lightweight neural operator approach for mean curvature motion to the Willmore flow. We train a single-time-step architecture with unconstrained and radial kernels that can be used for non-oriented interfaces in two and three dimensions, and provide an initialization method from 2D to 3D models.

Full work available at URL https://doi.org/10.48550/arXiv.2610.03534

Mathematics Subject Classification ID

Higher-order geometric flows (53E40) Applications of differential geometry to data and computer science (53Z50) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65M99) Artificial neural networks and deep learning (68T07)

arXiv classification

math.NA

Cites work

• A Cahn--Hilliard--Willmore phase field model for non-oriented interfaces
• A deep learning method for the dynamics of classic and conservative Allen-Cahn equations based on fully-discrete operators
• A hybrid minimizing movement and neural network approach to Willmore flow
• A phase field approach in the numerical study of the elastic bending energy for vesicle membranes
• A phase field based PDE constrained optimization approach to time discrete Willmore flow
• A singular perturbation problem with integral curvature bound
• An efficient iterative method for reconstructing surface from point clouds
• Asymptotic Analysis of Phase Field Formulations of Bending Elasticity Models
• Colliding interfaces in old and new diffuse-interface approximations of Willmore-flow
• Convergence of numerical approximations to a phase field bending elasticity model of membrane deformations
• DGM: a deep learning algorithm for solving partial differential equations
• Energy Dissipation Preserving Physics Informed Neural Network for Allen-Cahn Equations
• Energy-dissipative evolutionary deep operator neural networks
• Evolution of elastic curves in R^n: Existence and computation
• Fourier Neural Operator for Parametric Partial Differential Equations
• Generation and propagation of interfaces for reaction-diffusion equations
• Geometrical Evolution of Developed Interfaces
• Gradient flow for the Willmore functional
• Learning phase field mean curvature flows with neural networks
• Motion of multiple junctions: A level set approach
• Numerical analysis of physics-informed neural networks and related models in physics-informed machine learning
• Numerical analysis of physics-informed neural networks and related models in physics-informed machine learning
• On a modified conjecture of De Giorgi
• Phase field approximation for Plateau's problem: a curve geodesic distance penalty approach
• Phase-field approximation of the Willmore flow
• Phase-field approximations of the Willmore functional and flow
• Phase-field DeepONet: physics-informed deep operator neural network for fast simulations of pattern formation governed by gradient flows of free-energy functionals
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Propagation of fronts in a nonlinear fourth order equation
• Quasi-optimal error estimates for the mean curvature flow with a forcing term
• Removability of point singularities of Willmore surfaces
• Scientific Machine Learning through Physics-Informed Neural Networks: Where we are and What's next
• Solving Allen-Cahn and Cahn-Hilliard Equations using the Adaptive Physics Informed Neural Networks
• The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
• The Willmore functional
• Threshold dynamics for high order geometric motions

This page was built for publication: Learning the Willmore flow with a lightweight neural operator

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7374850)