Updated Program

Talk Abstract:

Over the past decade, machine learning researchers have developed
techniques to estimate the joint distribution of a set of variables
that span multiple related entities. These methods, often grouped
under the rubric "relational learning", include probabilistic
relational models, relational Markov networks, and relational
dependency networks. These techniques build on work in artificial
intelligence, statistics, databases, graph theory, and social network
analysis, and they are profoundly expanding the phenomena that we can
understand and predict. However, new frontiers await.

In this talk, I will briefly survey some recent work in learning
probabilistic models of relational data, and discuss several
applications of these techniques, including fraud detection in the
U.S. securities industry. I will argue that current techniques are
capable of learning only a subset of the knowledge needed by
practitioners in these domains, and that informing effective action
often requires a causal model. I will examine the open question of
whether relational representations make the problem of learning
causal models easier or harder, and present some reasons for optimism
that relational representations may be able to greatly improve our
ability to learn such models.

Bio:

David Jensen is Associate Professor of Computer Science and Director
of the Knowledge Discovery Laboratory at the University of
Massachusetts Amherst. He received his doctorate from Washington
University in 1992. From 1991 to 1995, he served as an analyst with
the Office of Technology Assessment, an agency of the United States
Congress. His research focuses on machine learning and knowledge
discovery in relational data, with applications to social network
analysis, web mining, and fraud detection. He serves on the program
committees of the International Conference on Machine Learning and
the International Conference on Knowledge Discovery and Data Mining,
and he serves on the ACM SIGKDD Executive Committee.