Physics-informed neural networks based on adaptive weighted loss functions for Hamilton-Jacobi equations

DOI10.3934/mbe.2022601OpenAlexW4294686794MaRDI QIDQ2694112

Yanyan Li

Publication date: 28 March 2023

Published in: Mathematical Biosciences and Engineering (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.3934/mbe.2022601

zbMATH Keywords

Hamilton-Jacobi equationsphysics-informed neural networksadaptive weighted optimizerthe logarithmic meanthe periodicity requirements

Mathematics Subject Classification ID

Artificial intelligence (68Txx)

Related Items (2)

Loss-attentional physics-informed neural networks ⋮ Variable separated physics-informed neural networks based on adaptive weighted loss functions for blood flow model

Uses Software

• DiffSharp
• AdaGrad
• Adam
• DeepXDE

Cites Work

• Unnamed Item
• Unnamed Item
• Directly solving the Hamilton-Jacobi equations by Hermite WENO schemes
• Overcoming the curse of dimensionality for some Hamilton-Jacobi partial differential equations via neural network architectures
• Affordable, entropy-consistent Euler flux functions. II: Entropy production at shocks
• On the limited memory BFGS method for large scale optimization
• Shape-from-shading, viscosity solutions and edges
• High-order semi-discrete central-upwind schemes for multi-dimensional Hamilton-Jacobi equations
• New high-resolution semi-discrete central schemes for Hamilton-Jacobi equations
• Hidden physics models: machine learning of nonlinear partial differential equations
• A model of wound-healing angiogenesis in soft tissue
• Cellular pattern formation during Dictyostelium aggregation
• Multiscale topology optimization using neural network surrogate models
• Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
• DGM: a deep learning algorithm for solving partial differential equations
• A third-order WENO scheme based on exponential polynomials for Hamilton-Jacobi equations
• Optimally weighted loss functions for solving PDEs with neural networks
• A Hermite method with a discontinuity sensor for Hamilton-Jacobi equations
• On some neural network architectures that can represent viscosity solutions of certain high dimensional Hamilton-Jacobi partial differential equations
• A ghost structure finite difference method for a fractional Fitzhugh-Nagumo monodomain model on moving irregular domain
• A two-stage physics-informed neural network method based on conserved quantities and applications in localized wave solutions
• When and why PINNs fail to train: a neural tangent kernel perspective
• Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
• The asymptotic profile of a dengue fever model on a periodically evolving domain
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Impact of bacterial hyperinfectivity on cholera epidemics in a spatially heterogeneous environment
• Discontinuous solutions of Hamilton-Jacobi equations on networks
• High-resolution semi-discrete Hermite central-upwind scheme for multidimensional Hamilton-Jacobi equations
• Mapped WENO and weighted power ENO reconstructions in semi-discrete central schemes for Hamilton-Jacobi equations
• High-Order Essentially Nonoscillatory Schemes for Hamilton–Jacobi Equations
• A Viscosity Solutions Approach to Shape-From-Shading
• High-Order Central WENO Schemes for Multidimensional Hamilton-Jacobi Equations
• Weighted ENO Schemes for Hamilton--Jacobi Equations
• High-Resolution Nonoscillatory Central Schemes for Hamilton--Jacobi Equations
• Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks
• Adaptive Deep Learning for High-Dimensional Hamilton--Jacobi--Bellman Equations
• Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
• DeepXDE: A Deep Learning Library for Solving Differential Equations
• Locally adaptive activation functions with slope recovery for deep and physics-informed neural networks

This page was built for publication: Physics-informed neural networks based on adaptive weighted loss functions for Hamilton-Jacobi equations