Releasing SQL query answers under differential privacy

PhD Qualifying Examination


Title: "Releasing SQL query answers under differential privacy"

by

Mr. Wei DONG


Abstract:

Since its introduction, differential privacy has attracted lots of 
interest in both academia and industry as it provides promising privacy 
protection for individual information in data analysis. At the same time, 
more and more sensitive information is collected by organizations, and 
these data are largely analyzed by people with SQL queries, which makes 
the problem of releasing SQL query answers under differential privacy 
important. In the past few years, we have seen a great number of efforts 
put into this topic. These works can be separated into three main types by 
the type of queries they focus on: queries over a single relation, 
sub-graph counting queries and multi-way join counting queries. In this 
paper, we first review background knowledge about differential privacy and 
introduce some mainstream differentially private mechanisms. Then, we 
survey the work related to answering SQL query answers under differential 
privacy in the above three types. Finally, we conclude the problems of 
current works and discuss some potential research topics.


Date:			Monday, 31 August 2020

Time:                  	2:00pm - 4:00pm

Zoom meeting:           https://hkust.zoom.us/j/99936952096

Committee Members:	Prof. Ke Yi (Supervisor)
  			Prof. Dimitris Papadias (Chairperson)
 			Dr. Qiong Luo
 			Prof. Raymond Wong


**** ALL are Welcome ****