☆ 4.7 Article

A priori and a posteriori error estimates for the Deep Ritz method applied to the Laplace and Stokes problem

JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS (2023)

Related references

Note: Only part of the references are listed.

Article Computer Science, Interdisciplinary Applications

A neural network multigrid solver for the Navier-Stokes equations

Nils Margenberg et al.

Summary: We present DNN-MG, a deep neural network multigrid solver for the instationary Navier-Stokes equations. The combination of a geometric multigrid solver and a recurrent neural network with memory improves computational efficiency. DNN-MG reduces computation time by solving on coarse levels with the multi-grid method and correcting interpolated solutions on fine levels using a compact neural network.

JOURNAL OF COMPUTATIONAL PHYSICS (2022)

Add to Collection

Article Mathematics, Applied

Estimates on the generalization error of physics-informed neural networks for approximating a class of inverse problems for PDEs

Siddhartha Mishra et al.

Summary: This article focuses on the application of physics-informed neural networks (PINNs) to approximating inverse problems for partial differential equations (PDEs). Specifically, it investigates data assimilation or unique continuation problems and provides rigorous estimates on the generalization error of PINNs for these problems. The study establishes an abstract framework and utilizes conditional stability estimates for the inverse problem to derive the estimate on PINN generalization error, thereby justifying the use of PINNs in this context.

IMA JOURNAL OF NUMERICAL ANALYSIS (2022)

Add to Collection

Article Mathematics, Applied

STRUCTURE PRESERVATION FOR THE DEEP NEURAL NETWORK MULTIGRID SOLVER

Nils Margenberg et al.

Summary: The simulation of partial differential equations is crucial in numerical analysis and various fields, with deep neural network-based techniques showing promise as an efficient alternative to existing methods like finite elements. By combining both approaches in approximating the Navier-Stokes equations, it was found that neural network corrections violated the divergence freedom of velocity field, prompting the investigation of penalty terms and a network learning the stream function to improve simulation fidelity. Experimental results showed that the stream function-based approach outperformed others, enhancing both divergence freedom and overall fidelity.

ELECTRONIC TRANSACTIONS ON NUMERICAL ANALYSIS (2022)

Add to Collection

Article Physics, Mathematical

Deep Nitsche Method: Deep Ritz Method with Essential Boundary Conditions

Yulei Liao et al.

Summary: The proposed method introduces a new approach to address essential boundary conditions encountered in deep learning-based numerical solvers for partial differential equations, without requiring significant extra computational costs. By utilizing Nitsche's variational formulation to deal with this difficulty, the method is consistent and error estimates in the energy norm have been proven, with illustration on several representative problems posed in at most 100 dimensions.

COMMUNICATIONS IN COMPUTATIONAL PHYSICS (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators

Lu Lu et al.

Summary: This study extends the capabilities of neural networks with the introduction of the deep operator network (DeepONet), which can be used to learn various operators, including explicit and implicit operators. Different formulations of the input function space were studied and their effect on generalization error for 16 diverse applications was examined.

NATURE MACHINE INTELLIGENCE (2021)

Add to Collection

Article Mathematics, Applied

DeepXDE: A Deep Learning Library for Solving Differential Equations

Lu Lu et al.

Summary: This article introduces an overview, implementation and applications of Physics-Informed Neural Networks (PINNs), along with a new residual-based adaptive refinement (RAR) method. By comparing with finite element methods, the advantages of PINNs and the versatile applications of the Python library DeepXDE are demonstrated. Overall, DeepXDE contributes to the more rapid development of the emerging scientific machine learning field.

SIAM REVIEW (2021)

Add to Collection

Article Engineering, Multidisciplinary

An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications

E. Samaniego et al.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2020)

Add to Collection

Article Computer Science, Information Systems

Artificial Neural Network Methods for the Solution of Second Order Boundary Value Problems

Cosmin Anitescu et al.

CMC-COMPUTERS MATERIALS & CONTINUA (2019)

Add to Collection

Article Computer Science, Interdisciplinary Applications

DGM: A deep learning algorithm for solving partial differential equations

Justin Sirignano et al.

JOURNAL OF COMPUTATIONAL PHYSICS (2018)

Add to Collection

Article Computer Science, Artificial Intelligence

A unified deep artificial neural network approach to partial differential equations in complex geometries

Jens Berg et al.

NEUROCOMPUTING (2018)

Add to Collection

Article Mathematics, Interdisciplinary Applications

A machine learning framework for data driven acceleration of computations of differential equations

Siddhartha Mishra

MATHEMATICS IN ENGINEERING (2018)

Add to Collection

Article Mathematics

The Deep Ritz Method: A Deep Learning-Based Numerical Algorithm for Solving Variational Problems

E. Weinan et al.

COMMUNICATIONS IN MATHEMATICS AND STATISTICS (2018)

Add to Collection

Article Engineering, Multidisciplinary

A posteriori error estimation for the fractional step theta discretization of the incompressible Navier-Stokes equations

Dominik Meidner et al.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2015)

Add to Collection

Article Mathematics, Applied

Variational localizations of the dual weighted residual estimator

Thomas Richter et al.

JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS (2015)

Add to Collection

Article Mathematics, Applied

A posteriori error estimation for finite element discretization of parameter identification problems

R Becker et al.

NUMERISCHE MATHEMATIK (2004)

Add to Collection

© Peeref 2019-2024. All rights reserved.