Volume 25, Issue 4
Model Reduction with Memory and the Machine Learning of Dynamical Systems

Chao Ma, Jianchun Wang & Weinan E

Commun. Comput. Phys., 25 (2019), pp. 947-962.

Published online: 2018-12

[An open-access article; the PDF is free to any online user.]

Preview Full PDF 158 2049
Export citation
  • Abstract

The well-known Mori-Zwanzig theory tells us that model reduction leads to memory effect. For a long time, modeling the memory effect accurately and efficiently has been an important but nearly impossible task in developing a good reduced model. In this work, we explore a natural analogy between recurrent neural networks and the Mori-Zwanzig formalism to establish a systematic approach for developing reduced models with memory. Two training models-a direct training model and a dynamically coupled training model-are proposed and compared. We apply these methods to the Kuramoto-Sivashinsky equation and the Navier-Stokes equation. Numerical experiments show that the proposed method can produce reduced model with good performance on both short-term prediction and long-term statistical properties.

  • Keywords

Model reduction Mori-Zwanzig recurrent neural networks.

  • AMS Subject Headings

35Q35 64P99

  • Copyright

COPYRIGHT: © Global Science Press

  • Email address
  • References
  • Hide All
    View All

@Article{CiCP-25-947, author = {}, title = {Model Reduction with Memory and the Machine Learning of Dynamical Systems}, journal = {Communications in Computational Physics}, year = {2018}, volume = {25}, number = {4}, pages = {947--962}, abstract = {

The well-known Mori-Zwanzig theory tells us that model reduction leads to memory effect. For a long time, modeling the memory effect accurately and efficiently has been an important but nearly impossible task in developing a good reduced model. In this work, we explore a natural analogy between recurrent neural networks and the Mori-Zwanzig formalism to establish a systematic approach for developing reduced models with memory. Two training models-a direct training model and a dynamically coupled training model-are proposed and compared. We apply these methods to the Kuramoto-Sivashinsky equation and the Navier-Stokes equation. Numerical experiments show that the proposed method can produce reduced model with good performance on both short-term prediction and long-term statistical properties.

}, issn = {1991-7120}, doi = {https://doi.org/10.4208/cicp.OA-2018-0269}, url = {http://global-sci.org/intro/article_detail/cicp/12885.html} }
Copy to clipboard
The citation has been copied to your clipboard