Coordinate Encoding on Linear Grids for Physics-Informed Neural Networks

TL;DR

This paper introduces a coordinate-encoding layer on linear grid cells for physics-informed neural networks, significantly improving training convergence speed and computational efficiency in solving PDEs.

Contribution

It proposes a novel coordinate-encoding approach with grid cells and spline interpolation to enhance PINN training and reduce computational costs.

Findings

Faster convergence in training PINNs.

Reduced computational cost due to axis-independent grid cells.

Stable and efficient model training demonstrated through numerical experiments.

Abstract

In solving partial differential equations (PDEs), machine learning utilizing physical laws has received considerable attention owing to advantages such as mesh-free solutions, unsupervised learning, and feasibility for solving high-dimensional problems. An effective approach is based on physics-informed neural networks (PINNs), which are based on deep neural networks known for their excellent performance in various academic and industrial applications. However, PINNs struggled with model training owing to significantly slow convergence because of a spectral bias problem. In this study, we propose a PINN-based method equipped with a coordinate-encoding layer on linear grid cells. The proposed method improves the training convergence speed by separating the local domains using grid cells. Moreover, it reduces the overall computational cost by using axis-independent linear grid cells. The method also achieves efficient and stable model training by adequately interpolating the encoded coordinates between grid points using natural cubic splines, which guarantees continuous derivative functions of the model computed for the loss functions. The results of numerical experiments demonstrate the effective performance and efficient training convergence speed of the proposed method.