Optimization of Physics-Informed Neural Networks for Solving the Nolinear Schrodinger Equation

被引：4

作者：

Chuprov, I. ^{[1
]}

Gao, Jiexing ^{[1
]}

Efremenko, D. ^{[1
]}

Kazakov, E. ^{[1
]}

Buzaev, F. ^{[1
]}

Zemlyakov, V. ^{[1
]}

机构：

[1] Huawei Technol Co Ltd, Labs 2012, Moscow Res Ctr, Moscow 121099, Russia

来源：

DOKLADY MATHEMATICS | 2023年 / 108卷 / SUPPL 2期

关键词：

Physics-Informed Neural Networks; nonlinear Schrodinger equation; nonlinear fiber optics; fine-tuning neural networks; DEEP LEARNING FRAMEWORK;

D O I：

10.1134/S1064562423701120

中图分类号：

O1 [数学];

学科分类号：

0701 ; 070101 ;

摘要：

Physics Informed Neural Networks (PINN) is a promising method for solving partial differential equations using machine learning. In this paper we consider the application of PINN to the nonlinear Schrodinger equation to describe the propagation of signal in an optical fibre. The factors determining the convergence of PINN from the physical point of view are investigated. Estimates of the convergence domain of the method in terms of fibre length and pulse energy are obtained. It is shown that the application sinusoidal activation function, as well as the weighting of the loss function terms are able to extend the convergence region of PINN with respect to fibre length and pulse energy. A generalization of the method (meta-PINN) is derived, allowing to solve the equation at its various parameters by using the pre-trained neural network.

引用

页码：S186 / S195

页数：10

共 23 条

[1]

Agrawal G. P., 2021, Fiber-Optic Communication Systems, V5th, DOI [10.1002/9781119737391, DOI 10.1002/9781119737391]

[2] Critical Investigation of Failure Modes in Physics-informed Neural Networks [J].

Basir, Shamsulhaq ;

Senocak, Inane .

AIAA SCITECH 2022 FORUM, 2022,

[3]

Chuprov I., 2022, ARXIV

[4]

Grossmann Tamara G., 2023, ARXIV

[5]

Häger C, 2018, IEEE INT SYMP INFO, P1590, DOI 10.1109/ISIT.2018.8437734

[6] Extended Physics-Informed Neural Networks (XPINNs): A Generalized Space-Time Domain Decomposition Based Deep Learning Framework for Nonlinear Partial Differential Equations [J].

Jagtap, Ameya D. ;

Karniadakis, George Em .

COMMUNICATIONS IN COMPUTATIONAL PHYSICS, 2020, 28 (05) :2002-2041

[7] Physics-Informed Neural Network for Optical Fiber Parameter Estimation From the Nonlinear Schrodinger Equation [J].

Jiang, Xiaotian ;

Wang, Danshi ;

Chen, Xue ;

Zhang, Min .

JOURNAL OF LIGHTWAVE TECHNOLOGY, 2022, 40 (21) :7095-7105

[8]

Jiang XT, 2021, 2021 OPTICAL FIBER COMMUNICATIONS CONFERENCE AND EXPOSITION (OFC)

[9] Physics-informed machine learning [J].

Karniadakis, George Em ;

Kevrekidis, Ioannis G. ;

Lu, Lu ;

Perdikaris, Paris ;

Wang, Sifan ;

Yang, Liu .

NATURE REVIEWS PHYSICS, 2021, 3 (06) :422-440

[10]

Kazakov E., 2023, PARALLEL COMPUTATION, DOI [10.1007/978-3-031-38864-4_10, DOI 10.1007/978-3-031-38864-4_10]

← 1 2 3 →