L0 Regularization and Empirical Bayes Frameworks for Robust Data Analysis
日期：2020/07/27 - 2020/07/27
学术讲座: L0 Regularization and Empirical Bayes Frameworks for Robust Data Analysis
主讲人: Dr. Jing Liu, University of Illinois at Urbana-Champaign
时间: 2020年7月27日 （周一）上午9点
讲座摘要
Being robust to outliers or malicious agents is of paramount importance when we learn from the big data. This talk will introduce a novel L0-regularization framework for outlier-robust data analysis. In the context of robust linear regression and robust PCA problems, I will show that the proposed framework can recover the underlying signal/subspace exactly in the noiseless setting, and stably in the noisy setting. Furthermore, we will see that this framework can tolerate more outliers and has smaller error bound than the L1-relaxation approach. I will briefly discuss the applications of this framework in robust matrix sensing, secure state estimation, and robust deep autoencoders. In the second part of my talk, I will discuss the Empirical Bayes framework for outlier-robust data analysis. Particularly, we will consider the sensor networks where some sensors are reliable, and the rest are unreliable. I will discuss the robustness and efficiency of the Empirical Bayes method in harvesting the information from those unreliable sensors in the presence of adversarial attacks. Lastly, I will conclude with the future extensions of this framework to distributed learning.
主讲人简介
Jing Liu (刘京) is a postdoctoral research associate of Coordinated Science Lab (CSL) in University of Illinois at Urbana-Champaign (UIUC). He obtained his Ph.D. degree from University of California at San Diego (UCSD) in 2019, under Prof. Bhaskar Rao. He received a M.S. in Electronic Engineering from Tsinghua University in 2013, and a B.E. in Electronic Engineering from Beijing Institute of Technology (BIT) in 2010. He received the first prize of Beijing Science & Technology Award, in 2013. He was the recipient of the National Fellowship of China, Guanghua Fellowship of Tsinghua University, Silver Medal of BIT, Frontiers of Innovation Fellowship of UCSD, and Shannon Graduate Fellowship nomination award of UCSD. His research interests include data science, Internet of Things (IoT), computer vision, and distributed learning.