PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics
The numerical solution of spatiotemporal partial differential equations (PDEs) using the deep learning method has attracted considerable attention in quantum mechanics, fluid mechanics, and many other natural sciences. In this paper, we propose an interactive temporal physics-informed neural network...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2024-12-01
|
| Series: | Electronic Research Archive |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/era.2024310 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832590782328995840 |
|---|---|
| author | Ruohan Cao Jin Su Jinqian Feng Qin Guo |
| author_facet | Ruohan Cao Jin Su Jinqian Feng Qin Guo |
| author_sort | Ruohan Cao |
| collection | DOAJ |
| description | The numerical solution of spatiotemporal partial differential equations (PDEs) using the deep learning method has attracted considerable attention in quantum mechanics, fluid mechanics, and many other natural sciences. In this paper, we propose an interactive temporal physics-informed neural network architecture based on ConvLSTM for solving spatiotemporal PDEs, in which the information feedback mechanism in learning is introduced between the current input and the previous state of network. Numerical experiments on four kinds of classical spatiotemporal PDEs tasks show that the extended models have superiority in accuracy, long-range learning ability, and robustness. Our key takeaway is that the proposed network architecture is capable of learning information correlation of the PDEs model with spatiotemporal data through the input state interaction process. Furthermore, our method also has a natural advantage in carrying out physical information and boundary conditions, which could improve interpretability and reduce the bias of numerical solutions. |
| format | Article |
| id | doaj-art-e36bb7714c0b416b82b4dae8f13ffb5b |
| institution | Kabale University |
| issn | 2688-1594 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Electronic Research Archive |
| spelling | doaj-art-e36bb7714c0b416b82b4dae8f13ffb5b2025-01-23T07:53:06ZengAIMS PressElectronic Research Archive2688-15942024-12-0132126641665910.3934/era.2024310PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamicsRuohan Cao0Jin Su1Jinqian Feng2Qin Guo3School of Science, Xi'an Polytechnic University, Xi'an 710048, ChinaSchool of Science, Xi'an Polytechnic University, Xi'an 710048, ChinaSchool of Science, Xi'an Polytechnic University, Xi'an 710048, ChinaSchool of Science, Xi'an Polytechnic University, Xi'an 710048, ChinaThe numerical solution of spatiotemporal partial differential equations (PDEs) using the deep learning method has attracted considerable attention in quantum mechanics, fluid mechanics, and many other natural sciences. In this paper, we propose an interactive temporal physics-informed neural network architecture based on ConvLSTM for solving spatiotemporal PDEs, in which the information feedback mechanism in learning is introduced between the current input and the previous state of network. Numerical experiments on four kinds of classical spatiotemporal PDEs tasks show that the extended models have superiority in accuracy, long-range learning ability, and robustness. Our key takeaway is that the proposed network architecture is capable of learning information correlation of the PDEs model with spatiotemporal data through the input state interaction process. Furthermore, our method also has a natural advantage in carrying out physical information and boundary conditions, which could improve interpretability and reduce the bias of numerical solutions.https://www.aimspress.com/article/doi/10.3934/era.2024310partial differential equationsphysics-informed deep learningdata-driven modelinginteractive learningneural network |
| spellingShingle | Ruohan Cao Jin Su Jinqian Feng Qin Guo PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics Electronic Research Archive partial differential equations physics-informed deep learning data-driven modeling interactive learning neural network |
| title | PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| title_full | PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| title_fullStr | PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| title_full_unstemmed | PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| title_short | PhyICNet: Physics-informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| title_sort | phyicnet physics informed interactive learning convolutional recurrent network for spatiotemporal dynamics |
| topic | partial differential equations physics-informed deep learning data-driven modeling interactive learning neural network |
| url | https://www.aimspress.com/article/doi/10.3934/era.2024310 |
| work_keys_str_mv | AT ruohancao phyicnetphysicsinformedinteractivelearningconvolutionalrecurrentnetworkforspatiotemporaldynamics AT jinsu phyicnetphysicsinformedinteractivelearningconvolutionalrecurrentnetworkforspatiotemporaldynamics AT jinqianfeng phyicnetphysicsinformedinteractivelearningconvolutionalrecurrentnetworkforspatiotemporaldynamics AT qinguo phyicnetphysicsinformedinteractivelearningconvolutionalrecurrentnetworkforspatiotemporaldynamics |