Speaker Name/Affiliation: William W. Cohen/Carneg

ie Mellon University

Talk Title: A Framework for Learning to Query

Heterogeneous Data

Date/Time: July 13 2006 at 1:00 p.m.

Loc

ation: ACES 2.402

Host: Raymond J. Mooney

Talk Abstract:
A long-term goal of research on data integration is to develop
data mo

dels and query languages that make it easy to answer
structured queries
using heterogeneous data. In this talk
I will describe a very simple

query language based on typed
similarity queries which are answered b

ased on a graph
containing a heterogeneous mixture of textual and non-t

extual
objects. The similarity metric proposed is based on a lazy
g

raph walk which can be approximated efficiently using
methods related

to particle filtering. Machine learning
techniques can be used to impro

ve this metric for specific
tasks often leading to performance far bet

ter than plausible
task-specific baseline methods. We experimentally ev

aluate
several classes of similarity queries from the domains of
an

alysis of biomedical text and personal information management:
for inst

ance in one set of experiments a user''s personal
information is repr

esented as a graph containing messages
calendar information social ne

twork information and a
timeline and similarity search is used to fin

d people likely
to attend a meeting.

This is joint work with Ein

at Minkov and Andrew Ng.

Speaker Bio:
William Cohen received hi

s bachelor''s degree in Computer
Science from Duke University in 1984

and a PhD in Computer
Science from Rutgers University in 1990. From 199

0 to 2000
Dr. Cohen worked at AT&T Bell Labs and later AT&T Labs-Resear

ch
and from April 2000 to May 2002 Dr. Cohen worked at Whizbang
La

bs a company specializing in extracting information from
the web. Dr.

Cohen is member of the board of the International
Machine Learning Soci

ety and has served as an action editor
for the Journal of Machine Lear

ning Research the journal
Machine Learning and the Journal of Artifici

al Intelligence
Research. He co-organized the 1994 International Machin

e
Learning Conference is the co-Program Committee Chair for
the 20

06 International Machine Learning Conference and
has served on more th

an 20 program committees or advisory
committees.

Dr. Cohen''s re

search interests include information integration
and machine learning

particularly information extraction
text categorization and learning f

rom large datasets. He
holds seven patents related to learning discove

ry information
retrieval and data integration and is the author of m

ore
than 100 publications.