Deep Hidden Physics Models: Deep Learning of Nonlinear Partial Differential Equations

A long-standing problem at the interface of artificial intelligence and applied mathematics is to devise an algorithm capable of achieving human level or even superhuman proficiency in transforming observed data into predictive mathematical models of the physical world. In the current era of abundance of data and advanced machine learning capabilities, the natural question arises: How can we automatically uncover the underlying laws of physics from high-dimensional data generated from experiments? In this work, we put forth a deep learning approach for discovering nonlinear partial differential equations from scattered and potentially noisy observations in space and time. Specifically, we approximate the unknown solution as well as the nonlinear dynamics by two deep neural networks. The first network acts as a prior on the unknown solution and essentially enables us to avoid numerical differentiations which are inherently ill-conditioned and unstable. The second network represents the nonlinear dynamics and helps us distill the mechanisms that govern the evolution of a given spatiotemporal data-set. We test the effectiveness of our approach for several benchmark problems spanning a number of scientific domains and demonstrate how the proposed framework can help us accurately learn the underlying dynamics and forecast future states of the system. In particular, we study the Burgers', Korteweg-de Vries (KdV), Kuramoto-Sivashinsky, nonlinear Schr\"{o}dinger, and Navier-Stokes equations.

• Publication year

  2018

• Venue

  Journal of machine learning research

• Publication date

  2018-01-20

• Fields of study

  Mathematics, Physics, Computer Science

• Identifiers
  arXiv 1801.06637
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Multistep Neural Networks for Data-driven Discovery of Nonlinear Dynamical Systems

  2018cited by this paper
• Forward-Backward Stochastic Neural Networks: Deep Learning of High-dimensional Partial Differential Equations

  2018cited by this paper
• Nonlinear information fusion algorithms for data-efficient multi-fidelity modelling

  2017cited by this paper
• Hidden physics models: Machine learning of nonlinear partial differential equations

  2017cited by this paper
• Automatic differentiation in PyTorch

  2017cited by this paper
• Parametric Gaussian process regression for big data

  2017cited by this paper
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations

  2017influential reference
• Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations

  2017influential reference
• Numerical Gaussian Processes for Time-Dependent and Nonlinear Partial Differential Equations

  2017cited by this paper
• Machine learning of linear differential equations using Gaussian processes

  2017cited by this paper
• Exact Recovery of Chaotic Systems from Highly Corrupted Data

  2016cited by this paper
• Deep Learning

  2016cited by this paper
• Inferring Biological Networks by Sparse Identification of Nonlinear Dynamics

  2016cited by this paper
• On the Expressive Power of Deep Neural Networks

  2016cited by this paper
• Deep Multi-fidelity Gaussian Processes

  2016cited by this paper
• Inferring solutions of differential equations using noisy multi-fidelity data

  2016cited by this paper
• Data-driven discovery of partial differential equations

  2016influential reference
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

  2016cited by this paper
• Chaos as an intermittently forced linear system

  2016cited by this paper
• Dynamic Mode Decomposition: Data-Driven Modeling of Complex Systems

  2016cited by this paper
• Equation-free mechanistic ecosystem forecasting using empirical dynamic modeling

  2015cited by this paper
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems

  2015cited by this paper
• Low-Dimensional Approach for Reconstruction of Airfoil Data via Compressive Sensing

  2015cited by this paper
• Automatic differentiation in machine learning: a survey

  2015influential reference
• Automated adaptive inference of phenomenological dynamical models

  2015cited by this paper
• On the Compressive Spectral Method

  2014cited by this paper
• Exploiting sparsity and equation-free architectures in complex systems

  2014cited by this paper
• Efficient Inference of Parsimonious Phenomenological Models of Cellular Dynamics Using S-Systems and Alternating Regression

  2014cited by this paper
• Analysis of Fluid Flows via Spectral Properties of the Koopman Operator

  2013cited by this paper
• Compressed modes for variational problems in mathematics and physics

  2013cited by this paper
• Compressive Sensing and Low-Rank Libraries for Classification of Bifurcation Regimes in Nonlinear Dynamical Systems

  2013cited by this paper
• Applied Koopmanism.

  2012cited by this paper
• Nonlinear Laplacian spectral analysis for time series with intermittency and low-frequency variability

  2012cited by this paper
• Sparse dynamics for partial differential equations

  2012cited by this paper
• Detecting Causality in Complex Ecosystems

  2012cited by this paper
• Automated refinement and inference of analytical models for metabolic networks

  2011cited by this paper
• Predicting catastrophes in nonlinear dynamical systems by compressive sensing.

  2011cited by this paper
• Normal forms for reduced stochastic climate models

  2009cited by this paper
• Distilling Free-Form Natural Laws from Experimental Data

  2009cited by this paper
• A fast immersed boundary method using a nullspace approach and multi-domain far-field boundary conditions

  2008cited by this paper
• The immersed boundary method: A projection approach

  2007cited by this paper
• Automated reverse engineering of nonlinear dynamical systems

  2007cited by this paper
• Spectral Properties of Dynamical Systems, Model Reduction and Decompositions

  2005cited by this paper
• Equation-Free, Coarse-Grained Multiscale Computation: Enabling Mocroscopic Simulators to Perform System-Level Analysis

  2003cited by this paper
• Amplitude equations from spatiotemporal binary-fluid convection data

  1999cited by this paper
• Identification of distributed parameter systems: A neural net based approach

  1998cited by this paper
• A comparison of recurrent training algorithms for time series analysis and system identification

  1996cited by this paper
• Regression Shrinkage and Selection via the Lasso

  1996cited by this paper
• DISCRETE- vs. CONTINUOUS-TIME NONLINEAR SIGNAL PROCESSING OF Cu ELECTRODISSOLUTION DATA

  1992cited by this paper
• Equations of Motion from a Data Series

  1987cited by this paper
• Spectral and finite difference solutions of the Burgers equation

  1986cited by this paper

• Multi-objective optimization of TBM-induced building settlement control considering physical constraints

  2026cites this paper
• Neuro-heuristic and regression-based modeling of three-dimensional Williamson-micropolar nanofluid flow with regularization and uncertainty quantification

  2026cites this paper
• Visualization of evaporation-driven flow inside an evaporating sessile droplet by means of Micro-PIV

  2026cites this paper
• Bayesian physics-informed neural networks with iterative ensemble Kalman inversion for RUL prediction and uncertainty quantification

  2026cites this paper
• Gradient-informed neural networks: Embedding prior beliefs for learning in low-data scenarios.

  2026cites this paper
• Error bounds for Physics Informed Neural Networks in Generalized KdV Equations placed on unbounded domains

  2026cites this paper
• An LLM-guided structural optimization method for aero-engine lubrication hydrocyclones

  2026cites this paper
• Latent selective state-space models for partial differential equations via progressive learning

  2026cites this paper
• A novel IDBO-VMD-iTransformer framework for air quality index prediction: Multi-strategy optimization and environmental sustainability assessment

  2026cites this paper
• Self-balancing physics-informed LSTM for soft-sensing of slurry concentration in dredger operation

  2026cites this paper
• Trend-guided deep physics-informed transfer learning method for remaining useful life prediction of rolling bearings across machines

  2026cites this paper
• A Parameterized Physics Informed Neural Network Solver for the Navier Stokes Equations Across Reynolds Numbers

  2026cites this paper
• A small-sample RUL prediction method based on physics-informed neural network with integrated domain-generalized features

  2026cites this paper
• The Ill-Posed Foundations of Physics-Informed Neural Networks and Their Finite-Difference Variants

  2026cites this paper
• PhysicsFormer: An Efficient and Fast Attention-Based Physics Informed Neural Network for Solving Incompressible Navier Stokes Equations

  2026cites this paper
• Advances and challenges in multi‑scale water environment system modeling: from process simulation to a novel simulator architecture

  2026cites this paper
• A data-driven model-free physical-informed deep operator network for solving nonlinear dynamic system

  2026cites this paper
• Physics-Informed Neural Networks for Constitutive Modeling and Multiphysics Coupling in Viscoelastic Materials: Applications to Asphalt Pavement Mechanics

  2026cites this paper
• Evolution of physics-informed neural networks: Recent architectural variants and optimization strategies

  2026cites this paper
• Adaptive Multi-Grade Deep Learning for Highly Oscillatory Fredholm Integral Equations of the Second Kind

  2026cites this paper
• Merging physics and neural network: A promising tool for prognostics and health management

  2026cites this paper
• Knowledge-data fusion oriented traffic state estimation: A stochastic physics-informed deep learning approach

  2026cites this paper
• A Physics-Informed Deep learning framework for traffic state estimation in signalized arterial roads

  2026cites this paper
• Deep surrogate model of groundwater and salt dynamics for optimizing subsurface pipes layout in irrigation areas

  2026cites this paper
• Analysis of traffic density dynamics under varied noise conditions using data-driven partial differential equations

  2026cites this paper
• Near-Earth Asteroid orbit determination with Physics-Informed Extreme Learning Machine

  2026cites this paper
• Tool wear monitoring and multi-step forecasting in deep hole boring based on physics-informed neural network and ARIMA model

  2026cites this paper
• A data-physics fusion network with multidimensional features for battery health prediction under small-sample and transfer scenarios

  2026cites this paper
• High-Fidelity Three-Dimensional Aerodynamic Flow Prediction on Wings with Physics-Constrained Dual-Parallel Attention UNet++

  2026cites this paper
• NewPINNs: Physics-Informing Neural Networks Using Conventional Solvers for Partial Differential Equations

  2026cites this paper
• A joint optimization approach to identifying sparse dynamics using least squares kernel collocation

  2025cites this paper
• When is a System Discoverable from Data? Discovery Requires Chaos

  2025cites this paper
• A physics-informed U-Net-LSTM network for nonlinear structural response under seismic excitation

  2025cites this paper
• Universal differential equations as a unifying modeling language for neuroscience

  2025cites this paper
• PINN Balls: Scaling Second-Order Methods for PINNs with Domain Decomposition and Adaptive Sampling

  2025influential citation
• Remaining useful life prediction using attention-based hidden physics-informed neural network and dual-path mixer model

  2025cites this paper
• A Physics-Informed Neural Network Combined with Domain Generalization Method for Harmonic Reducers Rul Prediction

  2025cites this paper
• Modal-integrated physics-informed neural networks for time-varying moving oscillator and beam interaction system

  2025cites this paper
• Direct method phase-based model identification for coupled oscillators

  2025cites this paper
• Physics-informed machine learning meets renewable energy systems: A review of advances, challenges, guidelines, and future outlooks

  2025cites this paper
• Data-driven solutions and parameter estimations for the integrable generalized fifth-order nonlinear Schrödinger equation based on physics-informed neural networks

  2025cites this paper
• Physics-Inspired Spatial Temporal Graph Neural Networks for Predicting Industrial Chain Resilience

  2025cites this paper
• Physical-Based Neural Network with Degradation Model for Remaining Useful Life Prediction of Aero-engine

  2025cites this paper
• Physics-Based Neural Network for Unsupervised Health Assessment of Lithium-Ion Batteries

  2025cites this paper
• A physics-constrained Bayesian neural network for machinery remaining useful life prediction and uncertainty quantification

  2025cites this paper
• Performance Improvement of Seismic Response Prediction Using the LSTM-PINN Hybrid Method

  2025cites this paper
• Learning with Mandelbrot and Julia

  2025cites this paper
• Robust Physics-Informed Neural Network for State of Health Estimation of Lithium-Ion Battery

  2025influential citation
• Joint prediction of SOH and RUL for Lithium-ion batteries by an enhanced Transformer model with physical information constraints

  2025cites this paper
• Variational Neural Networks for Observable Thermodynamics (V-NOTS)

  2025cites this paper
• Post-earthquake residual track irregularities prediction using wavelet-packet-transformation-powered domain adaptive neural network

  2025cites this paper
• Equation-Free Coarse Control of Distributed Parameter Systems via Local Neural Operators

  2025cites this paper
• Physics-Informed High-order Graph Dynamics Identification Learning for Predicting Complex Networks Long-term Dynamics

  2025cites this paper
• Energy-based methods for solving forward and inverse linear elasticity problems in 2D structures

  2025cites this paper
• Bead geometry prediction in wire arc directed energy deposition using physics-informed machine learning and low-fidelity data

  2025cites this paper
• Physics-informed neural networks for degradation diagnosis of lithium-ion batteries via electrochemical impedance spectroscopy

  2025cites this paper
• Development of Advanced Hybrid Modeling Techniques for Liquid Production Forecasting in Waterflooded Reservoirs Using Historical Injection and Production Data

  2025cites this paper
• Dual-driven prediction on stratum displacement induced by shield tunnel excavation based on machine learning and analytical methods

  2025cites this paper
• Integrated Framework of LSTM and Physical-Informed Neural Network for Lithium-Ion Battery Degradation Modeling and Prediction

  2025cites this paper
• Balance and Brighten: A Twin-Propeller Network to Release Potential of Physics Laws for Traffic State Estimation

  2025cites this paper
• Looks Great, Functions Better: Physics Compliance Text-to-3D Shape Generation

  2025cites this paper
• A Lightweight Hybrid Prediction Framework for Remaining Useful Life Based on CS-TCN-SAPINN

  2025influential citation
• Solved in Unit Domain: JacobiNet for Differentiable Coordinate Transformations

  2025cites this paper
• Thermal–hydraulic-mechanical-chemical-biological (THMCB) coupling in microbial induced carbonate precipitation (MICP): A comprehensive review

  2025cites this paper
• Research Progress and Technology Outlook of Deep Learning in Seepage Field Prediction During Oil and Gas Field Development

  2025cites this paper
• From clutter to clarity: Emergent neural operators via questionnaire metrics

  2025cites this paper
• A Sparse Bayesian Learning Algorithm for Estimation of Interaction Kernels in Motsch-Tadmor Model

  2025cites this paper
• Transformer-based Koopman autoencoder for linearizing Fisher's equation.

  2025cites this paper
• Predictive Digital Twins for Thermal Management Using Machine Learning and Reduced-Order Models

  2025cites this paper
• Artificial neural networks computing for heat transfer flow of hybrid nanofluid in rectangular geometry.

  2025cites this paper
• Predictive Modeling of Ball Mill Load Parameters Using Hybrid Physics-Informed Neural Networks

  2025cites this paper
• Foundation Models for Environmental Science: A Survey of Emerging Frontiers

  2025cites this paper
• Fractal-constrained deep learning for super-resolution of turbulence with zero or few label data

  2025cites this paper
• BPINN-EM-Post: Stochastic Electromigration Damage Analysis in the Post-Void Phase based on Bayesian Physics-Informed Neural Network

  2025cites this paper
• Observations on Recurrent Loss in the Neural Network Model of a Partial Differential Equation: the Advection-Diffusion Equation

  2025cites this paper
• Physics-Constrained Neural Network for Metasurface Optical Response Prediction

  2025cites this paper
• Interpretable Machine Learning in Physics: A Review

  2025cites this paper
• Development and performance analysis of a deep-learned finite element for 2D heat conduction on the natural coordinate system

  2025cites this paper
• Solving nonlinear Boussinesq equation of second-order time derivatives with physics-informed neural networks

  2025cites this paper
• Fast and Modular Whole-Body Lagrangian Dynamics of Legged Robots with Changing Morphology

  2025cites this paper
• Physically informed neural networks for homogenization and localization of composites with periodic microstructures

  2025cites this paper
• Physics-encoded convolutional attention network for forward and inverse analysis of spatial-temporal parabolic dynamics considering discontinuous heterogeneity

  2025cites this paper
• Integration of machine learning with finite element analysis in materials science: a review

  2025cites this paper
• On the definition and importance of interpretability in scientific machine learning

  2025cites this paper
• Three-dimensional sound field reconstruction from optical projections using physics-informed neural networks.

  2025cites this paper
• Function-on-function Differential Regression

  2025cites this paper
• Physics-informed neural network for prediction of scour depth using natural frequency of monopiles

  2025cites this paper
• Kirchhoff Integral Migration for VTI Media with Travel-Time Calculated Based on PINN

  2025cites this paper
• Estimation of Lithium Battery Health Status and Prediction of Remaining Useful Life Based on Improved PINN

  2025cites this paper
• Cost-effective Reduced-Order Modeling via Bayesian Active Learning

  2025cites this paper
• Small Data Explainer - The impact of small data methods in everyday life

  2025cites this paper
• A Statistician’s Overview of Physics-Informed Neural Networks for Spatio-Temporal Data

  2025cites this paper
• Mathematical Modeling of a pH Swing Precipitation Process and its Optimal Design

  2025influential citation
• Toward accurate RUL and SOH estimation using reinforced graph-based PINNs enhanced with dynamic weights

  2025cites this paper
• Neural Tangent Kernel of Neural Networks with Loss Informed by Differential Operators

  2025cites this paper
• A variational autoencoder based lightweight physics-informed neural network for remaining useful life prediction

  2025cites this paper
• Improving Long-term Autoregressive Spatiotemporal Predictions: A Proof of Concept with Fluid Dynamics

  2025cites this paper
• Physics-Informed Regression: Parameter Estimation in Parameter-Linear Nonlinear Dynamic Models

  2025cites this paper
• Finance Informed Neural Networks with Residual Gain Attention Based Radial Basis Functions

  2025cites this paper
• Review of empowering computer-aided engineering with artificial intelligence

  2025cites this paper