UTCS Colloquia/AI -Jennifer Neville/Purdue University, "How to learn from one sample? Statistical relational learning for single network", ACES 2.302

Contact Name: 
Jenna Whitney
Date: 
Nov 11, 2011 11:00am - 12:00pm

There is a sign-up schedule for this event that can be found at

http://apps.cs.utexas.edu/talkschedules/cgi/list_events.cgi

Type o

f Talk: UTCS Colloquia/Ai

Speaker/Affiliation: Jennifer Neville/Purdue
University

Talk Audience: UTCS Faculty, Grad Students, Undergrads a

nd Outside Interested Parties

Date/Time: Friday, November 11, 2011,
11:00 a.m.

Location: ACES 2.302

Host: Ray Mooney

Talk Title

: How to learn from one sample? Statistical relational learning for single

network

Talk Abstract:
Machine learning researchers focus on two dis

tinct learning scenarios for structured data that can be represented as a g

raph (i.e., where there are dependencies among the class labels and attrib

utes of linked nodes). In one scenario, the domain consists of a populatio

n of structured examples (e.g., chemical compounds). In this case, since

the population is comprised of a set of independent graphs, we can reason

about models and algorithms as the number of structured examples increases.
In the other scenario, the domain consists of a single, potentially infi

nite-sized network (e.g., the Facebook friendship network). In this case,
an increase in dataset size corresponds to acquiring a larger portion of t

he underlying network. In single network domains, even when there are mult

iple networks samples available for learning, they correspond to subnetwor

ks drawn from the same underlying network and thus may be dependent.
In o

ur recent work, we have focused on the development and analysis of statist

ical relational learning (SRL) methods for single network domains, particu

larly social networks. Although SRL algorithms have been successfully appli

ed for social network classification, the algorithmic foundations of SRL m

ethods are based on an implicit assumption of an underlying population of n

etworks---which does not hold for most social network datasets. In this tal

k, I will present our recent efforts to outline a more formal foundation f

or single network learning and discuss how the analysis has informed the de

velopment of more accurate estimation and evaluation methods.

Speaker

Bio:
Jennifer Neville is an assistant professor at Purdue University with
a joint appointment in the Departments of Computer Science and Statistics.
She received her PhD from the University of Massachusetts Amherst in 2006.
She received a DARPA IPTO Young Investigator Award in 2003 and was selecte

d as a member of the DARPA Computer Science Study Group in 2007. In 2008,

she was chosen by IEEE as one of "AI''s 10 to watch." Her research focuses

on developing data mining and machine learning techniques for relational do

mains, including citation analysis, fraud detection, and social network

analysis.