Scientific Machine Learning Through Physics–Informed Neural Networks: Where we are and What’s Next
Научное машинное обучение с использованием физически информированных нейронных сетей: современное состояние и дальнейшие перспективы
2022-07-26
SCID: 54.1/345pyarc
Discuss with AI
Collocation-based neural networksFinite Element MethodPartial differential equationsPhysics-constrained neural networksPhysics-informed neural networks
Figures from the paper
Abstract (AI)
Abstract Physics-Informed Neural Networks (PINN) are neural networks (NNs) that encode model equations, like Partial Differential Equations (PDE), as a component of the neural network itself. PINNs are nowadays used to solve PDEs, fractional equations, integral-differential equations, and stochastic PDEs. This novel methodology has arisen as a multi-task learning framework in which a NN must fit observed data while reducing a PDE residual. This article provides a comprehensive review of the literature on PINNs: while the primary goal of the study was to characterize these networks and their related advantages and disadvantages. The review also attempts to incorporate publications on a broader range of collocation-based physics informed neural networks, which stars form the vanilla PINN, as well as many other variants, such as physics-constrained neural networks (PCNN), variational hp-VPINN, and conservative PINN (CPINN). The study indicates that most research has focused on customizing the PINN through different activation functions, gradient optimization techniques, neural network structures, and loss function structures. Despite the wide range of applications for which PINNs have been used, by demonstrating their ability to be more feasible in some contexts than classical numerical techniques like Finite Element Method (FEM), advancements are still possible, most notably theoretical issues that remain unresolved.
Key Findings
1
Most PINN research focuses on modifying activation functions, optimization methods, neural architectures, and loss-function formulations.
2
PINNs can be more feasible than classical numerical methods such as the Finite Element Method in some applications, but important theoretical issues remain unresolved.
3
PINNs have been applied to PDEs, fractional equations, integro-differential equations, and stochastic PDEs.
4
Physics-Informed Neural Networks encode governing equations, such as PDEs, within neural networks and jointly fit observed data while minimizing equation residuals.
5
The review broadens coverage beyond vanilla PINNs to collocation-based variants including PCNN, variational hp-VPINN, and conservative PINN.
Research Object
Physics-Informed Neural Networks (PINNs) and their collocation-based variants (e.g., PCNN, hp-VPINN, CPINN)
Research Subject
the characteristics, advantages, disadvantages, application domains, customization strategies, and unresolved theoretical issues of PINNs
Publication Details
Publication Date
2022-07-26
Journal
Publisher
ISSN
Cited by
2819
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai9
Deep Reinforcement Learning: A Brief Survey2017
Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks2021
Review of Deep Learning Algorithms and Architectures2019
The rise of deep learning in drug discovery2018
Physics-Informed Neural Networks for Heat Transfer Problems2021
Toward Causal Representation Learning2021
Interpretable machine learning: Fundamental principles and 10 grand challenges2022
A Survey of Deep Learning Methods for Cyber Security2019
A-PINN: Auxiliary physics informed neural networks for forward and inverse problems of nonlinear integro-differential equations2022
Cited by6
Physics-informed machine learning in geotechnical engineering: a direction paper2025
Bridging urban theory and artificial intelligence: a multi-agent recommendation system for sustainable city development2026
Application of PINNs to Define Roughness Coefficients for Channel Flow Problems2025
Artificial Intelligence for Hydraulic-Fracturing Decision Support: A Workflow-Oriented Critical Review2026
Physics-Informed Neural Network for Reconstructing Free-Surface Transient Flow Fields in Long-Distance Water-Conveyance Tunnels from Sparse Observations2026
A survey on machine learning approaches for uncertainty quantification of engineering systems2025