Details concerning the starting workshop Conditional Independence Structures and Graphical Models

The precise description of conditional independence structures and the one-to-one correspondence of these structures with their visual or algebraic representation is fundamental to the development and implementation of algorithms for estimation, testing, etc. This description can be done using different types of graphs but also using non-graphical tools based on information theory such as imsets or matroids. The concept of conditional independence often occurs also in miscellaneous non-probabilistic calculi for dealing with uncertainty in artificial intelligence. The purpose of the seminar is to first recall the relatively simple mathematical tools needed for the axiomatic or graphical characterization of conditional independence structures (during the starting week's workshop). Then, researchers will focus on the open problems in this area mentioned in the original proposal of the seminar or raised by other participants of the workshop. This type of research can be viewed as the research for the precise mathematical description of all conditional independence structures, using both graphs and non-graphical methods.

MAIN SPEAKERS

D. R. Cox Nuffield College, Oxford
J. A. T. Koster Erasmus University, Rotterdam
M. D. Perlman University of Washington, Seattle
P. P. Shenoy Kansas University, Lawrence
R. W. Yeung Chinese University of Hong Kong
A. P. Dawid University College, London
A. Paz Technion, Haifa
G. Shafer Rutgers University, New jersey
N. Wermuth ZUMA Center for Survey Research, Mannheim
Z. Zhang University of Southern California, Los Angeles

INVITED SPEAKERS