Efficient SimRank Computation over Graphs

PhD Qualifying Examination


Title: "Efficient SimRank Computation over Graphs"

by

Mr. Yue WANG


Abstract:

Currently many real information systems can be modeled as graphs. Measuring 
similarity among nodes over graphs plays a key role in graph mining and 
analysis, and has many applications such as recommendation system, spam 
detection, graph clustering and link prediction. Among different node 
similarity measurements, SimRank is one of the most promising and popular ones. 
It can produce high quality results due to its recursive definition. However, 
the computational cost for SimRank is high, thus it has received a lot of 
research attention since introduced. Furthermore, real-world graphs evolve over 
time typically, which requires computing similarity scores efficiently over 
dynamic graphs. In this survey, we study current works for computing SimRank 
over both static and dynamic graphs, and compare the techniques used in 
different algorithms. The strengths and weaknesses of different methods are 
also discussed.


Date:			Thursday, 3 May 2018

Time:                  	1:00pm - 3:00pm

Venue:                  Room 2304
                         Lifts 17/18

Committee Members:	Prof. Lei Chen (Supervisor)
 			Dr. Yangqiu Song (Chairperson)
 			Dr. Raymond Wong
 			Dr. Ke Yi


**** ALL are Welcome ****