Abstract: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.

Multi-fidelity Physics-informed Generative Adversarial Network for Solving Partial Differential Equations

Feature-adjacent multi-fidelity physics-informed machine learning for partial differential equations

Physics-Informed Generator-Encoder Adversarial Networks with Latent Space Matching for Stochastic Differential Equations

Physics-Informed Generative Adversarial Network Solution to Buckley–Leverett Equation

Composing Partial Differential Equations with Physics-Aware Neural Networks

Diffusion-Generative Multi-Fidelity Learning for Physical Simulation

Physics-Informed Graph-Mesh Networks for PDEs: A hybrid approach for complex problems

Integration of physics-informed operator learning and finite element method for parametric learning of partial differential equations

Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations

Multi-physics Simulation Guided Generative Diffusion Models with Applications in Fluid and Heat Dynamics

Accelerated Solutions of Coupled Phase-Field Problems using Generative Adversarial Networks

An unsupervised latent/output physics-informed convolutional-LSTM network for solving partial differential equations using peridynamic differential operator

Multiphysics Design Optimization via Generative Adversarial Networks

Navier–stokes Generative Adversarial Network: a physics-informed deep learning model for fluid flow generation

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

MultiAdam: Parameter-wise Scale-invariant Optimizer for Multiscale Training of Physics-informed Neural Networks

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

Learning a Neural Solver for Parametric PDE to Enhance Physics-Informed Methods

LatentPINNs: Generative physics-informed neural networks via a latent representation learning

Data-driven forward-inverse problems of the 2-coupled mixed derivative nonlinear Schrödinger equation using deep learning