For more information about this meeting, contact Jason Morton.
|Title:||Algebraic aspects of conditional independence|
|Seminar:||Applied Algebra and Network Theory Seminar|
|Speaker:||Johannes Rauh, Max Planck Institute for Mathematics in the Sciences in Leipzig|
|Independence is an important statistical concept. From a theoretical point of view independence is the easiest
assumption that allows to make general statements about large numbers of random variables, as they appear in limit
theorems, large deviations and asymptotic statistics. On the other hand, it is the task of a statistician to find out whether
random variables are independent of each other, or conditionally independent, or how they depend on each other.
This idea leads to the theory of graphical models.
Conditional independence (CI) statements can be formulated as algebraic equations, and reasoning with CI statements can
be done by analyzing the corresponding CI ideals. Natural algebraic questions are to decide radicality or regularity of
these CI ideals. A primary decomposition of the CI ideal is useful to characterize the set of probability distributions
that satisfy the corresponding CI statements.
In my talk I plan to give an overview over the topic, including the CI axioms and recent radicality results for a class of CI ideals related to the study of robustness.|
Room Reservation Information
|Date:||04 / 25 / 2012|
|Time:||02:30pm - 03:20pm|