Object Identification over Structured and Unstructured Data

PhD Qualifying Examination


Title: "Object Identification over Structured and Unstructured Data"

Mr. Shaoxu Song


Abstract:

The essential goal of object identification is to return the objects in
the database that are duplicates or describing the same entities to the
given query in real world. With the emergence of various data from
heterogeneous sources, structure and unstructured, the definition of
objects turns to be various as well in databases. For example, an object
can be a collection of word tokens or attribute values, or a network of
tuples in a database. A more complex object may even consist of tuples
from heterogeneous databases.  In this survey, rather than providing a
longitudinal review of the object identification studies in decades of
years, we study the techniques in categories of identifying objects at
different data levels. Specifically, we discuss the different kinds of
techniques for the object identification at collection level, inside a
single database, and across heterogeneous data sources. Moreover, we also
present some of our previous work and future directions for the object
identification topics.


Date:     		Monday, 28 January 2008

Time:                   3:00p.m.-5:00p.m.

Venue:                  Room 3301A
			lifts 17-18

Committee Members:      Dr. Lei Chen (Supervisor)
			Dr. Wilfred Ng (Chairperson)
			Prof. Frederick Lochovsky
			Dr. Ke Yi


**** ALL are Welcome ****