skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

Abstract

Hejre, we introduce physics-informed neural networks – neural networks that are trained to solve supervised learning tasks while respecting any given laws of physics described by general nonlinear partial differential equations. In this work, we present our developments in the context of solving two main classes of problems: data-driven solution and data-driven discovery of partial differential equations. Depending on the nature and arrangement of the available data, we devise two distinct types of algorithms, namely continuous time and discrete time models. The first type of models forms a new family of data-efficient spatio-temporal function approximators, while the latter type allows the use of arbitrarily accurate implicit Runge–Kutta time stepping schemes with unlimited number of stages. Moreover, the effectiveness of the proposed framework is demonstrated through a collection of classical problems in fluids, quantum mechanics, reaction–diffusion systems, and the propagation of nonlinear shallow-water waves.

Authors:
ORCiD logo [1]; ORCiD logo [2];  [1]
  1. Brown Univ., Providence, RI (United States)
  2. Univ. of Pennsylvania, Philadelphia, PA (United States)
Publication Date:
Research Org.:
Univ. of Pennsylvania, Philadelphia, PA (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
OSTI Identifier:
1595805
Alternate Identifier(s):
OSTI ID: 1635941
Grant/Contract Number:  
SC0019116; N66001-15-2-4055; FA9550-17-1-0013
Resource Type:
Accepted Manuscript
Journal Name:
Journal of Computational Physics
Additional Journal Information:
Journal Volume: 378; Journal Issue: C; Related Information: https://github.com/PredictiveIntelligenceLab/PINNs; Journal ID: ISSN 0021-9991
Publisher:
Elsevier
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING; Data-driven scientific computing; Machine learning; Predictive modeling; Runge–Kutta methods; Nonlinear dynamics

Citation Formats

Raissi, Maziar, Perdikaris, Paris, and Karniadakis, George Em. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. United States: N. p., 2018. Web. doi:10.1016/j.jcp.2018.10.045.
Raissi, Maziar, Perdikaris, Paris, & Karniadakis, George Em. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. United States. doi:https://doi.org/10.1016/j.jcp.2018.10.045
Raissi, Maziar, Perdikaris, Paris, and Karniadakis, George Em. Sat . "Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations". United States. doi:https://doi.org/10.1016/j.jcp.2018.10.045. https://www.osti.gov/servlets/purl/1595805.
@article{osti_1595805,
title = {Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations},
author = {Raissi, Maziar and Perdikaris, Paris and Karniadakis, George Em},
abstractNote = {Hejre, we introduce physics-informed neural networks – neural networks that are trained to solve supervised learning tasks while respecting any given laws of physics described by general nonlinear partial differential equations. In this work, we present our developments in the context of solving two main classes of problems: data-driven solution and data-driven discovery of partial differential equations. Depending on the nature and arrangement of the available data, we devise two distinct types of algorithms, namely continuous time and discrete time models. The first type of models forms a new family of data-efficient spatio-temporal function approximators, while the latter type allows the use of arbitrarily accurate implicit Runge–Kutta time stepping schemes with unlimited number of stages. Moreover, the effectiveness of the proposed framework is demonstrated through a collection of classical problems in fluids, quantum mechanics, reaction–diffusion systems, and the propagation of nonlinear shallow-water waves.},
doi = {10.1016/j.jcp.2018.10.045},
journal = {Journal of Computational Physics},
number = C,
volume = 378,
place = {United States},
year = {2018},
month = {11}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 137 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Human-level concept learning through probabilistic program induction
journal, December 2015


Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning
journal, July 2015

  • Alipanahi, Babak; Delong, Andrew; Weirauch, Matthew T.
  • Nature Biotechnology, Vol. 33, Issue 8
  • DOI: 10.1038/nbt.3300

Inferring solutions of differential equations using noisy multi-fidelity data
journal, April 2017

  • Raissi, Maziar; Perdikaris, Paris; Karniadakis, George Em
  • Journal of Computational Physics, Vol. 335
  • DOI: 10.1016/j.jcp.2017.01.060

Machine learning of linear differential equations using Gaussian processes
journal, November 2017

  • Raissi, Maziar; Perdikaris, Paris; Karniadakis, George Em
  • Journal of Computational Physics, Vol. 348
  • DOI: 10.1016/j.jcp.2017.07.050

Multilayer feedforward networks are universal approximators
journal, January 1989


Data-driven discovery of partial differential equations
journal, April 2017

  • Rudy, Samuel H.; Brunton, Steven L.; Proctor, Joshua L.
  • Science Advances, Vol. 3, Issue 4
  • DOI: 10.1126/sciadv.1602614

Artificial neural networks for solving ordinary and partial differential equations
journal, January 1998

  • Lagaris, I. E.; Likas, A.; Fotiadis, D. I.
  • IEEE Transactions on Neural Networks, Vol. 9, Issue 5
  • DOI: 10.1109/72.712178

A hybrid neural network-first principles approach to process modeling
journal, October 1992


A paradigm for data-driven predictive modeling using field inversion and machine learning
journal, January 2016


Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
journal, October 2016

  • Ling, Julia; Kurzawski, Andrew; Templeton, Jeremy
  • Journal of Fluid Mechanics, Vol. 807
  • DOI: 10.1017/jfm.2016.615

Neural Network Modeling for Near Wall Turbulent Flow
journal, October 2002

  • Milano, Michele; Koumoutsakos, Petros
  • Journal of Computational Physics, Vol. 182, Issue 1
  • DOI: 10.1006/jcph.2002.7146

Multifidelity Information Fusion Algorithms for High-Dimensional Systems and Massive Data sets
journal, January 2016

  • Perdikaris, Paris; Venturi, Daniele; Karniadakis, George Em
  • SIAM Journal on Scientific Computing, Vol. 38, Issue 4
  • DOI: 10.1137/15M1055164

Why Does Deep and Cheap Learning Work So Well?
journal, July 2017

  • Lin, Henry W.; Tegmark, Max; Rolnick, David
  • Journal of Statistical Physics, Vol. 168, Issue 6
  • DOI: 10.1007/s10955-017-1836-5

Wavelet Scattering Regression of Quantum Chemical Energies
journal, January 2017

  • Hirn, Matthew; Mallat, Stéphane; Poilvert, Nicolas
  • Multiscale Modeling & Simulation, Vol. 15, Issue 2
  • DOI: 10.1137/16M1075454

Understanding deep convolutional networks
journal, April 2016

  • Mallat, Stéphane
  • Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 374, Issue 2065
  • DOI: 10.1098/rsta.2015.0203

On the limited memory BFGS method for large scale optimization
journal, August 1989

  • Liu, Dong C.; Nocedal, Jorge
  • Mathematical Programming, Vol. 45, Issue 1-3
  • DOI: 10.1007/BF01589116

Large Sample Properties of Simulations Using Latin Hypercube Sampling
journal, May 1987


Sparse grids
journal, May 2004


Discovering governing equations from data by sparse identification of nonlinear dynamical systems
journal, March 2016

  • Brunton, Steven L.; Proctor, Joshua L.; Kutz, J. Nathan
  • Proceedings of the National Academy of Sciences, Vol. 113, Issue 15
  • DOI: 10.1073/pnas.1517384113

    Works referencing / citing this record:

    Discovering Physical Concepts with Neural Networks
    journal, January 2020


    Prediction of aerodynamic flow fields using convolutional neural networks
    journal, June 2019


    Data‐Driven Materials Science: Status, Challenges, and Perspectives
    journal, September 2019

    • Himanen, Lauri; Geurts, Amber; Foster, Adam Stuart
    • Advanced Science, Vol. 6, Issue 21
    • DOI: 10.1002/advs.201900808

    General solutions for nonlinear differential equations: a rule-based self-learning approach using deep reinforcement learning
    journal, May 2019


    Simulation of nonlinear fractional dynamics arising in the modeling of cognitive decision making using a new fractional neural network
    journal, December 2019

    • Hadian Rasanan, Amir Hosein; Bajalan, Nastaran; Parand, Kourosh
    • Mathematical Methods in the Applied Sciences, Vol. 43, Issue 3
    • DOI: 10.1002/mma.5981

    Solving Fokker-Planck equation using deep learning
    journal, January 2020

    • Xu, Yong; Zhang, Hao; Li, Yongge
    • Chaos: An Interdisciplinary Journal of Nonlinear Science, Vol. 30, Issue 1
    • DOI: 10.1063/1.5132840

    Massive computational acceleration by using neural networks to emulate mechanism-based biological models
    journal, September 2019


    Interactive design and variation of hull shapes: pros and cons of different CAD approaches
    journal, September 2019

    • Vernengo, Giuliano; Villa, Diego; Gaggero, Stefano
    • International Journal on Interactive Design and Manufacturing (IJIDeM), Vol. 14, Issue 1
    • DOI: 10.1007/s12008-019-00613-3

    Prediction of aerodynamic flow fields using convolutional neural networks
    journal, June 2019


    Multiscale Modeling Meets Machine Learning: What Can We Learn?
    journal, February 2020

    • Peng, Grace C. Y.; Alber, Mark; Buganza Tepole, Adrian
    • Archives of Computational Methods in Engineering
    • DOI: 10.1007/s11831-020-09405-5

    Neural network models for the anisotropic Reynolds stress tensor in turbulent channel flow
    journal, December 2019


    Modal Analysis of Fluid Flows: Applications and Outlook
    journal, March 2020

    • Taira, Kunihiko; Hemati, Maziar S.; Brunton, Steven L.
    • AIAA Journal, Vol. 58, Issue 3
    • DOI: 10.2514/1.j058462

    Integrating machine learning and multiscale modeling—perspectives, challenges, and opportunities in the biological, biomedical, and behavioral sciences
    journal, November 2019

    • Alber, Mark; Buganza Tepole, Adrian; Cannon, William R.
    • npj Digital Medicine, Vol. 2, Issue 1
    • DOI: 10.1038/s41746-019-0193-y

    Deep neural network learning of complex binary sorption equilibria from molecular simulation data
    journal, January 2019

    • Sun, Yangzesheng; DeJaco, Robert F.; Siepmann, J. Ilja
    • Chemical Science, Vol. 10, Issue 16
    • DOI: 10.1039/c8sc05340e

    Memory embedded non-intrusive reduced order modeling of non-ergodic flows
    journal, December 2019

    • Ahmed, Shady E.; Rahman, Sk. Mashfiqur; San, Omer
    • Physics of Fluids, Vol. 31, Issue 12
    • DOI: 10.1063/1.5128374

    Nonintrusive reduced order modeling framework for quasigeostrophic turbulence
    journal, November 2019


    Machine learning and artificial intelligence to aid climate change research and preparedness
    journal, November 2019

    • Huntingford, Chris; Jeffers, Elizabeth S.; Bonsall, Michael B.
    • Environmental Research Letters, Vol. 14, Issue 12
    • DOI: 10.1088/1748-9326/ab4e55

    Solving partial differential equations by a supervised learning technique, applied for the reaction–diffusion equation
    journal, November 2019


    Neural networks catching up with finite differences in solving partial differential equations in higher dimensions
    journal, January 2020


    Uniformly accurate machine learning-based hydrodynamic models for kinetic equations
    journal, October 2019

    • Han, Jiequn; Ma, Chao; Ma, Zheng
    • Proceedings of the National Academy of Sciences, Vol. 116, Issue 44
    • DOI: 10.1073/pnas.1909854116

    A priori analysis on deep learning of subgrid-scale parameterizations for Kraichnan turbulence
    journal, January 2020

    • Pawar, Suraj; San, Omer; Rasheed, Adil
    • Theoretical and Computational Fluid Dynamics, Vol. 34, Issue 4
    • DOI: 10.1007/s00162-019-00512-z

    A deep learning enabler for nonintrusive reduced order modeling of fluid flows
    journal, August 2019

    • Pawar, S.; Rahman, S. M.; Vaddireddy, H.
    • Physics of Fluids, Vol. 31, Issue 8
    • DOI: 10.1063/1.5113494

    Discovering a universal variable-order fractional model for turbulent Couette flow using a physics-informed neural network
    journal, December 2019

    • Mehta, Pavan Pranjivan; Pang, Guofei; Song, Fangying
    • Fractional Calculus and Applied Analysis, Vol. 22, Issue 6
    • DOI: 10.1515/fca-2019-0086

    Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
    journal, January 2020

    • Raissi, Maziar; Yazdani, Alireza; Karniadakis, George Em
    • Science, Vol. 367, Issue 6481
    • DOI: 10.1126/science.aaw4741

    Feature engineering and symbolic regression methods for detecting hidden physics from sparse sensor observation data
    journal, January 2020

    • Vaddireddy, Harsha; Rasheed, Adil; Staples, Anne E.
    • Physics of Fluids, Vol. 32, Issue 1
    • DOI: 10.1063/1.5136351