Speaker: 姜立建教授，同济大学

Inviter: 于海军(AMSS,研究员), 翟佳羽(上海科技大学), 廖奇峰(上海科技大学)

Title: Transfer Operators and Transport Maps for Bayesian Filtering

Language: Chinese

Time & Venue: 2026.10.26 15:10-16:10 腾讯会议：602-613-835

Abstract: Bayesian filtering is essential for state estimation and uncertainty quantification in stochastic dynamical systems. In this talk, I will present several approaches to nonlinear Bayesian filtering based on transfer operators and probability transport. First, we introduce a Perron–Frobenius operator filter for characterizing probability distribution evolution and Bayesian updates. Next, we develop an ensemble transport filter based on maximum mean discrepancy (MMD) optimization. Finally, we present a coupling-informed transport filtering framework that exploits the dependence between state and observation variables. These methods provide complementary perspectives and computational tools for nonlinear and non-Gaussian Bayesian filtering.