An Analysis of Physics-Informed Neural Networks

Edward Small

2023-03-06

Abstract:Whilst the partial differential equations that govern the dynamics of our world have been studied in great depth for centuries, solving them for complex, high-dimensional conditions and domains still presents an incredibly large mathematical and computational challenge. Analytical methods can be cumbersome to utilise, and numerical methods can lead to errors and inaccuracies. On top of this, sometimes we lack the information or knowledge to pose the problem well enough to apply these kinds of methods. Here, we present a new approach to approximating the solution to physical systems - physics-informed neural networks. The concept of artificial neural networks is introduced, the objective function is defined, and optimisation strategies are discussed. The partial differential equation is then included as a constraint in the loss function for the optimisation problem, giving the network access to knowledge of the dynamics of the physical system it is modelling. Some intuitive examples are displayed, and more complex applications are considered to showcase the power of physics informed neural networks, such as in seismic imaging. Solution error is analysed, and suggestions are made to improve convergence and/or solution precision. Problems and limitations are also touched upon in the conclusions, as well as some thoughts as to where physics informed neural networks are most useful, and where they could go next.

Machine Learning,Dynamical Systems,Numerical Analysis

What problem does this paper attempt to address?

The paper attempts to address the significant mathematical and computational challenges in solving partial differential equations (PDEs) under complex and high-dimensional conditions. Traditional analytical methods may be difficult to apply, while numerical methods may lead to errors and inaccuracies. Additionally, in some cases, we may lack sufficient information or knowledge to properly define the problem. Therefore, the paper proposes a new approach—Physics-Informed Neural Networks (PINNs)—to approximate the solutions of physical systems.

Specifically, the paper first introduces the basic concepts of artificial neural networks, defines the objective function, and discusses optimization strategies. Then, it incorporates partial differential equations as constraints in the loss function of the optimization problem, enabling the neural network to leverage the knowledge of the dynamics of the modeled physical system. By presenting some intuitive examples and more complex applications (such as seismic imaging), the paper demonstrates the powerful capabilities of Physics-Informed Neural Networks. The paper also analyzes the errors in the solutions and offers suggestions for improving convergence speed and accuracy, while also discussing the issues and limitations of the method. Finally, the paper explores the most suitable fields for Physics-Informed Neural Networks and their future development directions.