Building 22 Room 127 and ONLINE via ZOOM (Meeting ID : 894 1270 2994) (Passcode : 3GQbG7) Link: https://us02web.zoom.us/j/89412702994?pwd=L1ZyZjBSODRLa0hsRUVUVU1JWVptdz09
Nov. 4, 2021 - Nov. 4, 2021
10 a.m. - 11 a.m.
Speaker:
Dr. Muhammad Emzir
Assistant Professor
Control and Instrumentation Engineering Department
King Fahd University of Petroleum and Minerals
Dhahran, Saudi Arabia
Abstract:
In this seminar, I will talk about some of the recent works that I do in the field of stochastic filtering. In particular, I will talk about projection filters. The projection filter approximates the dynamic of conditional probability densities for the optimal filtering problems. In projection filters, the stochastic partial differential equation corresponds to the evolution of the optimal filtering densities is projected to a manifold of parametric densities, yielding a finite-dimensional stochastic differential equation. Despite its capability of capturing complex probability density shapes, the implementations of projection filters are (so far) restricted to either the Gaussian family or unidimensional filtering problems. In this talk, I will discuss a new combination of numerical integration and auto-differentiation to solve the projection filter. Moreover, I will also cover some expositions about this proposal for the manifold of the exponential family. Lastly, I will present some numerical experiments to show that this approach can maintain a fairly accurate approximation of the filtering density compared to the finite-difference scheme solutions for the Zakai filter and the particle filter with a relatively low number of quadrature points.
Speaker Bio :
Dr. Muhammad Emzir obtained his PhD degree from the University of New South Wales, Australia in 2018. In 2017, he joined the Australian National University, as a postdoctoral researcher. From 2018 to August, 2021, he was a postdoctoral researcher at Aalto University. From August 2021, he joined King Fahd University of Petroleum and Minerals as an assistant professor at Control and Instrumentation Engineering Department. His research interests are stochastic filter, nonlinear control, quantum filtering and control.
All faculty, researchers and graduate students are invited to attend.