用于物理信息神经网络的 DiffGrad

Jamshaid Ul Rahman, Nimra
{"title":"用于物理信息神经网络的 DiffGrad","authors":"Jamshaid Ul Rahman, Nimra","doi":"arxiv-2409.03239","DOIUrl":null,"url":null,"abstract":"Physics-Informed Neural Networks (PINNs) are regarded as state-of-the-art\ntools for addressing highly nonlinear problems based on partial differential\nequations. Despite their broad range of applications, PINNs encounter several\nperformance challenges, including issues related to efficiency, minimization of\ncomputational cost, and enhancement of accuracy. Burgers' equation, a\nfundamental equation in fluid dynamics that is extensively used in PINNs,\nprovides flexible results with the Adam optimizer that does not account for\npast gradients. This paper introduces a novel strategy for solving Burgers'\nequation by incorporating DiffGrad with PINNs, a method that leverages the\ndifference between current and immediately preceding gradients to enhance\nperformance. A comprehensive computational analysis is conducted using\noptimizers such as Adam, Adamax, RMSprop, and DiffGrad to evaluate and compare\ntheir effectiveness. Our approach includes visualizing the solutions over space\nat various time intervals to demonstrate the accuracy of the network. The\nresults show that DiffGrad not only improves the accuracy of the solution but\nalso reduces training time compared to the other optimizers.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"124 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"DiffGrad for Physics-Informed Neural Networks\",\"authors\":\"Jamshaid Ul Rahman, Nimra\",\"doi\":\"arxiv-2409.03239\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Physics-Informed Neural Networks (PINNs) are regarded as state-of-the-art\\ntools for addressing highly nonlinear problems based on partial differential\\nequations. Despite their broad range of applications, PINNs encounter several\\nperformance challenges, including issues related to efficiency, minimization of\\ncomputational cost, and enhancement of accuracy. Burgers' equation, a\\nfundamental equation in fluid dynamics that is extensively used in PINNs,\\nprovides flexible results with the Adam optimizer that does not account for\\npast gradients. This paper introduces a novel strategy for solving Burgers'\\nequation by incorporating DiffGrad with PINNs, a method that leverages the\\ndifference between current and immediately preceding gradients to enhance\\nperformance. A comprehensive computational analysis is conducted using\\noptimizers such as Adam, Adamax, RMSprop, and DiffGrad to evaluate and compare\\ntheir effectiveness. Our approach includes visualizing the solutions over space\\nat various time intervals to demonstrate the accuracy of the network. The\\nresults show that DiffGrad not only improves the accuracy of the solution but\\nalso reduces training time compared to the other optimizers.\",\"PeriodicalId\":501369,\"journal\":{\"name\":\"arXiv - PHYS - Computational Physics\",\"volume\":\"124 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Computational Physics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.03239\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Computational Physics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.03239","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

物理信息神经网络(PINNs)被认为是解决基于偏微分方程的高度非线性问题的最先进工具。尽管 PINNs 的应用范围很广,但它也遇到了一些性能挑战,包括与效率、计算成本最小化和提高精度有关的问题。布尔格斯方程是流体力学中的一个基本方程,在 PINNs 中得到了广泛应用,其亚当优化器可提供灵活的结果,但不考虑过去的梯度。本文通过将 DiffGrad 与 PINNs 结合在一起,介绍了一种求解伯格斯方程的新策略,这种方法利用当前梯度与紧接其后的梯度之间的差异来提高性能。本文使用 Adam、Adamax、RMSprop 和 DiffGrad 等优化器进行了全面的计算分析,以评估和比较它们的有效性。我们的方法包括将不同时间间隔的空间解决方案可视化,以展示网络的准确性。结果表明,与其他优化器相比,DiffGrad 不仅提高了解决方案的准确性,还缩短了训练时间。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
DiffGrad for Physics-Informed Neural Networks
Physics-Informed Neural Networks (PINNs) are regarded as state-of-the-art tools for addressing highly nonlinear problems based on partial differential equations. Despite their broad range of applications, PINNs encounter several performance challenges, including issues related to efficiency, minimization of computational cost, and enhancement of accuracy. Burgers' equation, a fundamental equation in fluid dynamics that is extensively used in PINNs, provides flexible results with the Adam optimizer that does not account for past gradients. This paper introduces a novel strategy for solving Burgers' equation by incorporating DiffGrad with PINNs, a method that leverages the difference between current and immediately preceding gradients to enhance performance. A comprehensive computational analysis is conducted using optimizers such as Adam, Adamax, RMSprop, and DiffGrad to evaluate and compare their effectiveness. Our approach includes visualizing the solutions over space at various time intervals to demonstrate the accuracy of the network. The results show that DiffGrad not only improves the accuracy of the solution but also reduces training time compared to the other optimizers.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Designing a minimal Landau theory to stabilize desired quasicrystals Uncovering liquid-substrate fluctuation effects on crystal growth and disordered hyperuniformity of two-dimensional materials Exascale Quantum Mechanical Simulations: Navigating the Shifting Sands of Hardware and Software Influence of dislocations in multilayer graphene stacks: A phase field crystal study AHKASH: a new Hybrid particle-in-cell code for simulations of astrophysical collisionless plasma
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1