UTCS Colloquium/AI: Brian Milch/MIT: Probabilistic Reasoning about Large or Unknown Worlds ACES 2.402 Friday August 15 2008 11:00 a.m.

Contact Name: 
Jenna Whitney
Date: 
Aug 15, 2008 11:00am - 12:00pm

There is a signup schedule for this event (UT EID required).

Typ

e of Talk: UTCS Colloquim/AI

Speaker/Affiliation: Brian Milch/MIT<

br>
Date/Time: Friday August 15 2008 11:00 a.m.

Location: AC

ES 2.402

Host: Ray Mooney

Talk Title: Probabilistic Reasoni

ng about Large or Unknown Worlds

Talk Abstract:
Intelligent syste

ms must reason about the objects in their world
such as the vehicles on
a road monitored by video cameras or
the people and events mentioned i

n a set of documents. In many
cases these objects are not known in adva

nce; the system must
infer their existence. Reasoning about such large

and potentially
unknown worlds calls for probabilistic techniques that g

eneralize
across objects rather than handling each one individually. I w

ill
begin this talk by introducing Bayesian logic or BLOG a languagefor defining distributions over possible worlds with varying sets of
o

bjects. Every well-formed BLOG model is guaranteed to fully
define a dis

tribution even if it defines infinitely many random
variables. I will t

hen describe an approximate inference algorithm
for BLOG that executes a
Markov chain over partial worlds each
of which instantiates only a rel

evant subset of the variables. This
algorithm will be demonstrated on a

citation-matching problem:
identifying the distinct publications referre

d to by citation strings in
online papers. I will also describe an exact
inference algorithm
that is applicable to models with large but known s

ets of objects.
By exploiting two distinct forms of symmetry in such mod

els this
algorithm can achieve exponential speed-ups even compared toprevious lifted algorithms.

Speaker Bio:
Brian Milch recently c

ompleted a post-doc with Prof. Leslie
Kaelbling in the Computer Science

and Artificial Intelligence
Lab at MIT; he will join the Search Quality
group at Google in
September. Joining Google will be a homecoming for B

rian
since he also spent a year there after graduating from Stanford
in 2000. He then entered the Ph.D. program at U.C. Berkeley
where he w

orked with Prof. Stuart Russell and received his
Ph.D. in 2006. Brian is
the recipient of an NSF Graduate
Research Fellowship and a Siebel Schol

arship and was
named one of the Ten to Watch in artificial intelligence
by
IEEE Intelligent Systems in 2008.