Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers

Kiwon Um, Robert Brand, Yun, Fei, Philipp Holl +1 more
6/30/2020
physics.comp-phcs.LG

Official code: tum-pbs/Solver-in-the-Loop (180 stars, MIT, last commit 2022-10-19).

Abstract

Finding accurate solutions to partial differential equations (PDEs) is a crucial task in all scientific and engineering disciplines. It has recently been shown that machine learning methods can improve the solution accuracy by correcting for effects not captured by the discretized PDE. We target the problem of reducing numerical errors of iterative PDE solvers and compare different learning approaches for finding complex correction functions. We find that previously used learning approaches are significantly outperformed by methods that integrate the solver into the training loop and thereby allow the model to interact with the PDE during training. This provides the model with realistic input distributions that take previous corrections into account, yielding improvements in accuracy with stable rollouts of several hundred recurrent evaluation steps and surpassing even tailored supervised variants. We highlight the performance of the differentiable physics networks for a wide variety of PDEs, from non-linear advection-diffusion systems to three-dimensional Navier-Stokes flows.

View on arXivView PDF

Code Implementations(1)

tum-pbs/Solver-in-the-LoopOfficial100%
MIT

Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers

18033Python, MakefileJun 20, 20204 years agoMIT

Similar Papers

Text2PDE: Latent Diffusion Models for Accessible Physics Simulation

86% match1 repo402024

LIGO-PINN: Learned Initialization via Gated Optimization to Alleviate Convergence Failures in Physics Informed Neural Networks

86% match1 repo12026

Building Trust in PINNs: Error Estimation through Finite Difference Methods

86% match2026

Transferable Physics-Informed Representations via Closed-Form Head Adaptation

84% match2026

Physics-Informed Residuals for Adaptive Mesh Refinement in Finite-Difference PDE Solvers

84% match2026

Unsupervised Legendre-Galerkin Neural Network for Singularly Perturbed Partial Differential Equations

84% match2022

Cite this paper

@article{um2020solverintheloop,
  title  = {Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers},
  author = {Kiwon Um and Robert Brand and Yun and Fei and Philipp Holl and Nils Thuerey},
  year   = {2020},
  eprint = {2007.00016},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2007.00016}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2007.00016.svg)](https://codeofpaper.com/paper/2007.00016)

Discussion