Recent advance in machine learning for partial differential equation

被引：0
作者：
Ka Chun Cheung
Simon See
机构：
[1] NVIDIA Corp,NVIDIA AI Technology Centre
来源：
CCF Transactions on High Performance Computing | 2021年 / 3卷
关键词：
Machine learning; Partial differential equations; Physics informed neural network; Fourier neural operator;
D O I：
暂无
中图分类号：
学科分类号：
摘要：
Machine learning method has been applied to solve different kind of problems in different areas due to the great success in several tasks such as computer vision, natural language processing and robotic in recent year. In scientific computing community, it is well-known that solving partial differential equations, which are naturally derived from physical rules that describe some of phenomena, is a challenging task in terms of computational efficiency and model accuracy. On the other hand, machine learning models are data-driven that purely reply on learning the pattern of the data distribution. Researcher recently proposed a few new frameworks to solve certain kind of partial differential equations with machine learning technique. In this paper, we discuss two newly developed machine learning based methods for solving partial differential equations.
引用
收藏
页码：298 / 310
页数：12
相关论文
共 68 条
  • [11] Forgy EW(2018)Solving high-dimensional partial differential equations using deep learning Proc. Natl. Acad. Sci. 115 8505-8510
  • [12] González-García R(2020)Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks J. Comput. Phys. 404 109120-243
  • [13] Rico-Martínez R(1991)Nonlinear principal component analysis using autoassociative neural networks AIChE J. 37 233-1000
  • [14] Kevrekidis IG(1998)Artificial neural networks for solving ordinary and partial differential equations IEEE Trans. Neural Netw. 9 987-133
  • [15] Haber E(2020)Variational training of neural network approximations of solution maps for physical models J. Comput. Phys. 409 109338-955
  • [16] Ruthotto L(1943)A logical calculus of the ideas immanent in nervous activity Bull. Math. Biophys. 5 115-707
  • [17] Hayati M(2017)Sympy: symbolic computing in python PeerJ Comput. Sci. 3 e103-1364
  • [18] Karami B(2018)Deep hidden physics models: deep learning of nonlinear partial differential equations J. Mach. Learn. Res. 19 932-380
  • [19] Hermann J(2019)Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations J. Comput. Phys. 378 686-A665
  • [20] Schätzle Z(2018)DGM: a deep learning algorithm for solving partial differential equations J. Comput. Phys. 375 1339-undefined
1234567