Combining Tuple and Attribute Uncertainty in Probabilistic Databases

There has been a long history of work in the database communityon probabilistic databases. There is also a long tradition of work within
the machine learning and reasoning under uncertainty communities on
tractable factored representations of probability distributions, such as
probabilistic graphical models. Recently, research from these two commu-
nities are starting to share more and more commonalities, as the proba-
bilistic database work incorporates richer probabilistic dependencies and
the machine learning work incorporates richer relational models.
In this talk, I will survey some of the recent work, including work on
probabilistic relational models. I will show how these models can capture
both tuple and attribute uncertainty, and discuss effective query methods.
Joint work with Prithviraj Sen and Amol Deshpande.