Power-enhanced residual network for function approximation and physics-informed inverse problems

DOI10.1016/J.AMC.2024.128910MaRDI QIDQ6994316zbMATH OpenFDO

Authors Amir Noorizadegan, D. L. Young, Y. C. Hon, C. S. Chen

Publication date 24 April 2025

Published in Applied Mathematics and Computation (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/J.AMC.2024.128910

zbMATH Keywords

interpolationinverse problemneural networkspartial differential equationsfunction approximationresidual networkhighway network for PINN

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Numerical interpolation (65D05) Numerical radial basis function approximation (65D12)

Cites work

• scientific article; zbMATH DE number 1999247 (Why is no real title available?)
• On the search of the shape parameter in radial basis functions using univariate global optimization methods
• On the selection of a better radial basis function and its shape parameter in interpolation problems
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Scattered Data Interpolation: Tests of Some Method
• Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks

Cited in

(3)

• Stable weight updating: a key to reliable PDE solutions using deep learning
• On the convergence of PINNs for inverse source problem in the complex Ginzburg-Landau equation
• Physics-informed neural fractional differential equations

This page was built for publication: Power-enhanced residual network for function approximation and physics-informed inverse problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6994316)