Multi-scale time-stepping of partial differential equations with transformers

DOI10.1016/J.CMA.2024.116983MaRDI QIDQ6550140zbMATH OpenFDO

Authors AmirPouya Hemmasian, Amir Barati Farimani

Publication date 4 June 2024

Published in Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)

zbMATH Keywords

machine learningpartial differential equationstransformersnumerical time-stepping

Mathematics Subject Classification ID

Numerical optimization and variational techniques (65K10) Probabilistic models, generic numerical methods in probability and statistics (65C20) Multigrid methods; domain decomposition for initial value and initial-boundary value problems involving PDEs (65M55) Probabilistic methods, particle methods, etc. for initial value and initial-boundary value problems involving PDEs (65M75) Spectral, collocation and related methods for initial value and initial-boundary value problems involving PDEs (65M70)

Recommendations

• Transformers as neural operators for solutions of differential equations with finite regularity
• Local neural operator for solving transient partial differential equations on varied domains
• Time integration schemes based on neural networks for solving partial differential equations on coarse grids
• Accelerating explicit time-stepping with spatially variable time steps through machine learning

Cites work

• Data-driven science and engineering. Machine learning, dynamical systems, and control
• Machine learning for fluid mechanics
• Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders
• Non-intrusive reduced order modeling of nonlinear problems using neural networks
• Nonlinear mode decomposition with convolutional neural networks for fluid dynamics
• POD-based reduced-order models with deforming grids
• PhyGeoNet: physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Prediction of aerodynamic flow fields using convolutional neural networks
• Scientific machine learning through physics-informed neural networks: where we are and what's next
• Solving parametric PDE problems with artificial neural networks
• Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data
• The Random Feature Model for Input-Output Maps between Banach Spaces
• Transformers for modeling physical systems
• fPINNs: Fractional Physics-Informed Neural Networks

Cited in

(2)

• Physics-informed geometry-aware neural operator
• Transformers as neural operators for solutions of differential equations with finite regularity

This page was built for publication: Multi-scale time-stepping of partial differential equations with transformers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6550140)