pith. sign in

arxiv: 2207.00556 · v2 · pith:DRA3ATQZnew · submitted 2022-07-01 · 💻 cs.LG · physics.flu-dyn

Learning to correct spectral methods for simulating turbulent flows

Gideon Dresdner , Dmitrii Kochkov , Peter Norgaard , Leonardo Zepeda-N\'u\~nez , Jamie A. Smith , Michael P. Brenner , Stephan Hoyer This is my paper
classification 💻 cs.LG physics.flu-dyn
keywords numericalpdeslearningspectralmachinemethodsworkclassical
0
0 comments X
read the original abstract

Despite their ubiquity throughout science and engineering, only a handful of partial differential equations (PDEs) have analytical, or closed-form solutions. This motivates a vast amount of classical work on numerical simulation of PDEs and more recently, a whirlwind of research into data-driven techniques leveraging machine learning (ML). A recent line of work indicates that a hybrid of classical numerical techniques and machine learning can offer significant improvements over either approach alone. In this work, we show that the choice of the numerical scheme is crucial when incorporating physics-based priors. We build upon Fourier-based spectral methods, which are known to be more efficient than other numerical schemes for simulating PDEs with smooth and periodic solutions. Specifically, we develop ML-augmented spectral solvers for three common PDEs of fluid dynamics. Our models are more accurate (2-4x) than standard spectral solvers at the same resolution but have longer overall runtimes (~2x), due to the additional runtime cost of the neural network component. We also demonstrate a handful of key design principles for combining machine learning and numerical methods for solving PDEs.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. ANTIC: Adaptive Neural Temporal In-situ Compressor

    cs.LG 2026-04 unverdicted novelty 6.0

    ANTIC reduces storage for large-scale PDE simulations by orders of magnitude through adaptive temporal snapshot selection combined with continual neural-field residual compression while preserving physics accuracy.

  2. ANTIC: Adaptive Neural Temporal In-situ Compressor

    cs.LG 2026-04 unverdicted novelty 5.0

    ANTIC combines an adaptive temporal selector with continual neural field compression to reduce storage for high-dimensional PDE simulations by orders of magnitude while preserving physics accuracy.

  3. ANTIC: Adaptive Neural Temporal In-situ Compressor

    cs.LG 2026-04 unverdicted novelty 5.0

    ANTIC achieves orders-of-magnitude storage reduction for high-dimensional spatiotemporal PDE simulations by combining adaptive temporal snapshot selection with neural-field residual compression while preserving physic...

  4. Algorithmically Designed Artificial Neural Networks (ADANNs): Higher order deep operator learning for parametric partial differential equations

    math.NA 2023-02 unverdicted novelty 5.0

    ADANNs design ANN architectures and initializations to mimic classical numerical algorithms for parametric PDE operator approximation and report significant outperformance over existing methods in numerical tests.

  5. Integral regularization PINNs for evolution equations

    math.NA 2025-03 unverdicted novelty 4.0

    IR-PINNs improve long-time accuracy for evolution equations by enforcing integral constraints over time sub-intervals and using adaptive collocation point sampling.