\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENT SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, June 4, 2002}
\centerline{Sequoia Hall Room 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Daphne Koller}
\centerline{\sl Stanford University}
\bigskip
\centerline{\bf Learning Probabilistic Models from Relational Data}
\bigskip

Bayesian networks are a compact and natural representation for complex
probabilistic models.  They use graphical notation to encode domain
structure: the direct probabilistic dependencies between variables in the
domain.  However, many real-world domains are best described by relational
in which instances of multiple types are related to each other
in complex ways.  For example, in a scientific paper domain, papers are
related to each other via citation, and are also related to their authors.
Bayesian networks are attribute-based, making it difficult to represent
the rich relational structure of complex domains involving multiple
entities that interact with each other.  The talk will describe
probabilistic relational models (PRMs), a new probabilistic modeling
language suitable for relational domains.  PRMs extend the language of
Bayesian networks with the expressive power of object-relational
languages.  They model the uncertainty over the attributes of objects
in the domain as well as uncertainty over the existence of relations
between objects.  The talk will present techniques for automatically
inducing PRMs directly from a relational data set, and applications of
these techniques to pattern discovery, including relational classification
and cluster.  It will demonstrate the applicability of the techniques
on complex real-world data sets, including web data and genomic data.

Joint work with Pieter Abbeel, Nir Friedman, Lise Getoor, Avi Pfeffer,
Eran Segal, and Ben Taskar.

\bye