Zhibing Zhao (赵志冰) Ph.D.

Microsoft, Bellevue, WA

zhaozb08 AT gmail DOT com

My CV and Google Scholar


Short Bio

Zhibing Zhao got his Ph.D. in Computer Science at Rensselaer Polytechnic Institute (RPI). He received his master’s degree from the University of Connecticut and his bachelor’s degree from Tsinghua University, both in Electrical Engineering. His research focuses on efficient learning (statistical inference) of ranking models, in particular, the Random Utility Models (including the Plackett-Luce model) and their mixtures. His strength lies in variance characterization and complexity analysis of inference algorithms, and designing new algorithms that are both statistically and computationally efficient.


Rank aggregation problems have a wide range of applications, including presidential elections, yelp rating, college ranking, and so on. People (agents) have diverse preferences over the alternatives (presidential candidates, restaurants, etc.) even though there is a ground truth. Ranking models are used to model this uncertainty. My research focuses on efficient estimation of ground truth parameters of ranking models, mainly from the computational perspective. Specifically, I am interested in identifiability of a class of ranking models, the convergence rate of an algorithm, and the statistical accuracy.