期刊
NATURE COMMUNICATIONS
卷 12, 期 1, 页码 -出版社
NATURE PORTFOLIO
DOI: 10.1038/s41467-021-21331-z
关键词
-
资金
- NDSEG fellowship
- Air Force Office of Scientific Research [AFOSR FA9550-18-1-0200, AFOSR FA9550-17-1-0329]
- Washington Research Foundation
- Army Research Office [ARO W911NF-19-1-0045]
- Defense Advanced Research Projects Agency [DARPA PA-18-01-FP-125]
Traditional physics-based modeling relies on approximating observed dynamics as a balance between dominant processes within asymptotic regimes, but researchers have proposed a new approach using equation space to identify neglected terms in non-asymptotic regimes. Their data-driven balance models successfully delineate dominant physics in systems like turbulence, combustion, nonlinear optics, geophysical fluids, and neuroscience.
Throughout the history of science, physics-based modeling has relied on judiciously approximating observed dynamics as a balance between a few dominant processes. However, this traditional approach is mathematically cumbersome and only applies in asymptotic regimes where there is a strict separation of scales in the physics. Here, we automate and generalize this approach to non-asymptotic regimes by introducing the idea of an equation space, in which different local balances appear as distinct subspace clusters. Unsupervised learning can then automatically identify regions where groups of terms may be neglected. We show that our data-driven balance models successfully delineate dominant balance physics in a much richer class of systems. In particular, this approach uncovers key mechanistic models in turbulence, combustion, nonlinear optics, geophysical fluids, and neuroscience. The dynamics of complex physical systems can be determined by the balance of a few dominant processes. Callaham et al. propose a machine learning approach for the identification of dominant regimes from experimental or numerical data with examples from turbulence, optics, neuroscience, and combustion.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据