Xê-mi-na khoa học
23/06/2021: ThS. Nguyễn Văn Lý (San Diego State Univ., US), Fisher Information Neural Estimation
Fisher information is a well-known and well-defined concept in mathematical statistics. There are many areas in which the Fisher information is applied to, e.g., estimation theory, Bayesian statistics, frequentist statistics, optimal experimental design, computational neuroscience, physical laws, and machine learning. Therefore, the estimation of Fisher information is of critical importance. In this talk, we introduce a machine learning-based Fisher information estimation method, referred to as Fisher Information Neural Estimation (FINE). Most existing methods for Fisher information estimation rely on the estimation of the underlying distribution, which is not always possible. The proposed FINE method directly estimates the Fisher information based on the observed data. We also show via numerical examples that the proposed FINE method not only outperforms an existing method but also has a lower computational complexity.
Speaker: ThS. Nguyễn Văn Lý, San Diego State Univ.
Time: 15:30, Wednesday, June 23, 2021
Venue: Webinar; Access code: https://bit.ly/3zKq0W2
Nguyen Van Ly received the B.Eng. degree in Electronics and Telecommunications from the University of Engineering and Technology, Vietnam National University, Hanoi, Vietnam in 2014, and the M.Sc. degree in Advanced Wireless Communications Systems from CentraleSupélec, University of Paris-Saclay, France in 2016. Since August 2017, he has been a Ph.D. student in a joint doctoral program in computational science between San Diego State University and University of California, Irvine, CA, USA. His research interests include wireless communications, signal processing, and machine learning.