Mastering high-dimensional dynamics with Hamiltonian neural networks

Autor:	Miller, Scott T., Lindner, John F., Choudhary, Anshul, Sinha, Sudeshna, Ditto, William L.
Rok vydání:	2020
Předmět:	Computer Science - Neural and Evolutionary Computing Nonlinear Sciences - Chaotic Dynamics
Druh dokumentu:	Working Paper
Popis:	We detail how incorporating physics into neural network design can significantly improve the learning and forecasting of dynamical systems, even nonlinear systems of many dimensions. A map building perspective elucidates the superiority of Hamiltonian neural networks over conventional neural networks. The results clarify the critical relation between data, dimension, and neural network learning performance. Comment: 7 pages, 9 figures
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2008.04214 Zobrazit plný text záznamu View this record from Arxiv