Asif Hamid, Danish Rafiq, Shahkar Ahmad Nahvi, Mohammad Abid Bazaz
{"title":"利用坐标和流图深度学习提高多尺度系统的计算效率","authors":"Asif Hamid, Danish Rafiq, Shahkar Ahmad Nahvi, Mohammad Abid Bazaz","doi":"arxiv-2407.00011","DOIUrl":null,"url":null,"abstract":"Complex systems often show macroscopic coherent behavior due to the\ninteractions of microscopic agents like molecules, cells, or individuals in a\npopulation with their environment. However, simulating such systems poses\nseveral computational challenges during simulation as the underlying dynamics\nvary and span wide spatiotemporal scales of interest. To capture the\nfast-evolving features, finer time steps are required while ensuring that the\nsimulation time is long enough to capture the slow-scale behavior, making the\nanalyses computationally unmanageable. This paper showcases how deep learning\ntechniques can be used to develop a precise time-stepping approach for\nmultiscale systems using the joint discovery of coordinates and flow maps.\nWhile the former allows us to represent the multiscale dynamics on a\nrepresentative basis, the latter enables the iterative time-stepping estimation\nof the reduced variables. The resulting framework achieves state-of-the-art\npredictive accuracy while incurring lesser computational costs. We demonstrate\nthis ability of the proposed scheme on the large-scale Fitzhugh Nagumo neuron\nmodel and the 1D Kuramoto-Sivashinsky equation in the chaotic regime.","PeriodicalId":501167,"journal":{"name":"arXiv - PHYS - Chaotic Dynamics","volume":"216 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Enhancing Computational Efficiency in Multiscale Systems Using Deep Learning of Coordinates and Flow Maps\",\"authors\":\"Asif Hamid, Danish Rafiq, Shahkar Ahmad Nahvi, Mohammad Abid Bazaz\",\"doi\":\"arxiv-2407.00011\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Complex systems often show macroscopic coherent behavior due to the\\ninteractions of microscopic agents like molecules, cells, or individuals in a\\npopulation with their environment. However, simulating such systems poses\\nseveral computational challenges during simulation as the underlying dynamics\\nvary and span wide spatiotemporal scales of interest. To capture the\\nfast-evolving features, finer time steps are required while ensuring that the\\nsimulation time is long enough to capture the slow-scale behavior, making the\\nanalyses computationally unmanageable. This paper showcases how deep learning\\ntechniques can be used to develop a precise time-stepping approach for\\nmultiscale systems using the joint discovery of coordinates and flow maps.\\nWhile the former allows us to represent the multiscale dynamics on a\\nrepresentative basis, the latter enables the iterative time-stepping estimation\\nof the reduced variables. The resulting framework achieves state-of-the-art\\npredictive accuracy while incurring lesser computational costs. We demonstrate\\nthis ability of the proposed scheme on the large-scale Fitzhugh Nagumo neuron\\nmodel and the 1D Kuramoto-Sivashinsky equation in the chaotic regime.\",\"PeriodicalId\":501167,\"journal\":{\"name\":\"arXiv - PHYS - Chaotic Dynamics\",\"volume\":\"216 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-04-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Chaotic Dynamics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2407.00011\",\"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 - Chaotic Dynamics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2407.00011","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Enhancing Computational Efficiency in Multiscale Systems Using Deep Learning of Coordinates and Flow Maps
Complex systems often show macroscopic coherent behavior due to the
interactions of microscopic agents like molecules, cells, or individuals in a
population with their environment. However, simulating such systems poses
several computational challenges during simulation as the underlying dynamics
vary and span wide spatiotemporal scales of interest. To capture the
fast-evolving features, finer time steps are required while ensuring that the
simulation time is long enough to capture the slow-scale behavior, making the
analyses computationally unmanageable. This paper showcases how deep learning
techniques can be used to develop a precise time-stepping approach for
multiscale systems using the joint discovery of coordinates and flow maps.
While the former allows us to represent the multiscale dynamics on a
representative basis, the latter enables the iterative time-stepping estimation
of the reduced variables. The resulting framework achieves state-of-the-art
predictive accuracy while incurring lesser computational costs. We demonstrate
this ability of the proposed scheme on the large-scale Fitzhugh Nagumo neuron
model and the 1D Kuramoto-Sivashinsky equation in the chaotic regime.