主講人介紹：
Yunpeng Li is a Lecturer in Artificial Intelligence in the Department of Computer Science at the University of Surrey, U.K. He received the B.A. and M.Sc. degrees from the Beijing University of Posts and Telecommunications, China, in 2009 and 2012, respectively, and the Ph.D. degree in the Department of Electrical and Computer Engineering at McGill University, Canada in 2017. From 2017 to 2018, he was a Postdoctoral Research Assistant in Machine Learning at the Machine Learning Research Group, Department of Engineering Science at the University of Oxford, U.K. He was a Junior Research Fellow at the Wolfson College, University of Oxford in 2018. His research interests include Bayesian inference, Monte Carlo methods, object tracking, and statistical machine learning.
內容摘要：
Going with flow: transport methods in sequential Monte Carlo methods
Sequential state estimation in non-linear and non-Gaussian state spaces has a wide range of applications in signal processing and statistics. One of the most effective non-linear filtering approaches, particle filters a.k.a. sequential Monte Carlo methods, suffer from weight degeneracy in high-dimensional filtering scenarios. Several avenues have been pursued to address high dimensionality. Among these, particle flow filters migrate particles continuously from the prior distribution to the posterior distribution by solving partial differential equations. Approximations are needed in the implementation of all of these filters; as a result, the particles do not exactly match a sample drawn from the desired posterior distribution.
In this talk, I will present new filters which incorporate deterministic particle flows into an encompassing particle filter framework. The valuable theoretical guarantees concerning particle filter performance still apply, but we can exploit the attractive performance of the particle flow methods. The filters I will describe involve a computationally efficient weight update step, arising because the embedded particle flows we designed possess an invertible mapping property. I will demonstrate the advantage of the proposed particle flow particle filters through numerical simulations of a challenging multi-sensor multi-target tracking scenario.
