Seminar| Institute of Mathematical Sciences
Time: Friday, August 16th, 2024 , 10:00-11:00
Location：IMS S408
Speaker: Linjie Wen (闻林杰)，中国石油大学（北京）
Abstrct: Data assimilation is an essential task in the field of earth science in dealing with mathematical models of dynamical systems, and the filtering algorithm for nonlinear dynamical systems is a research hotspot in the field of data assimilation. In the first work we propose for modest dimensional filtering problems a defensive marginal PF algorithm which constructs a sampling distribution in the marginal space by combining the standard PF and the EnKF approximation using a multiple importance sampling (MIS) scheme. In the second work we propose transport mapping based variational ensemble Kalman filter for sequential Bayesian filtering problems with generic observation models, including affine-mapping and SVGD based mapping.