Recent Advances in Data Assimilation with Machine Learning
|
Organizer(s): |
Name:
|
Affiliation:
|
Country:
|
Nan Chen
|
University of Wisconsin-Madison
|
USA
|
Jinlong Wu
|
University of Wisconsin-Madison
|
0
|
Yeyu Zhang
|
Shanghai University of Finance and Economics
|
Peoples Rep of China
|
|
|
|
|
|
|
|
|
Introduction:
| | In recent years, machine learning has become one of the dominant methods for studying complex dynamical systems, especially for data assimilation and prediction. Machine learning has been widely used as a computationally efficient surrogate of complicated knowledge-based forecast models or applied to the role of a statistical correction for the knowledge-based models that mitigate the model error in the forecast step of data assimilation. It has also been exploited to optimize the tuning parameters in ensemble data assimilation, such as the inflation rate of the covariance matrix. On the other hand, machine learning has been applied more directly by building end-to-end learning schemes for the entire data assimilation pipeline. This special session will focus on topics that relate to both fundamental mathematical theories and numerical algorithms for data assimilation with the assistance of machine learning. The session serves as a venue for developing new ideas and advancing mathematical analysis of machine-learning-assisted data assimilation methods.
|
|
|
|
| |
