[IEEE-bhpjobs] Colloquium: Alex Ihler - Wednesday March 8, 2006

[Walter Heger]

Duke Computer Science Colloquia

Graphical Models for Estimation in Sensor Networks
Alex Ihler
D106 LSRC, Duke
Wednesday  March 8, 2006
11:45AM - 12:45PM


Abstract

Ubiquitous networks of sensors have the ability to provide new windows
into the environment through unprecedented volumes of observations and
data.  Making use of this data, however, poses new challenges in
information processing.  Graphical models (including Bayes' nets and
Markov random fields) have shown themselves to be a powerful framework
for dealing with learning and estimation problems in sensor networks,
allowing one to succinctly describe structure among the many random
variables of interest, organize computations efficiently, and develop
accurate approximations.

In this talk I will describe some of the ways in which graphical
models can be used to solve learning and estimation problems in sensor
networks, using two primary examples.  In the first half, I'll discuss
how the common initial task of self-localization, or estimating the
locations of the sensors given a set of noisy measurements, can be
elegantly framed as inference in a graphical model and solved using a
variant of belief propagation.  In the second half I'll describe a
machine learning task in a small sensor network, specifically creating
models of typical (human) behavior observed through the sensors and
using these models to detect the occurrence of unusual (anomalous)
events.


Biography

Alexander Ihler received his B.S. degree in Math and Electrical
Engineering from  the California Institute of Technology in 1998, and
S.M. and Ph.D. degrees in Electrical Engineering and Computer Science
from the Massachusetts Institute of Technology in 2000 and 2005 as a
member of the Stochastic Systems Group.  He is currently a
Postdoctoral Scholar at the University of California, Irvine.  His
research interests include statistical signal processing, machine
learning, and nonparametric statistics, with applications including
computational biology, atmospheric science, and sensor networks.

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/