**Fabio Cozman ^{1}
Escola Politécnica, University of São Paulo
fgcozman@usp.br --
http://www.cs.cmu.edu/~fgcozman/home.html**

This paper analyzes irrelevance and independence relations in graphical
models associated with convex sets of probability distributions (called
Quasi-Bayesian networks).
The basic question in Quasi-Bayesian networks is, How can
irrelevance/independence relations in Quasi-Bayesian networks be detected,
enforced and exploited? This paper addresses these questions through
Walley's definitions of irrelevance and independence.
Novel algorithms and results are presented for inferences with
the so-called natural extensions using fractional linear programming,
and the properties of the so-called type-1 extensions are
clarified through a new generalization of d-separation.

- INTRODUCTION
- BACKGROUND MATERIAL
- LOCALLY DEFINED QUASI-BAYESIAN NETWORKS
- TYPE-1 EXTENSION
- NATURAL EXTENSION
- EXAMPLE
- CONCLUSION
- PROOFS
- Acknowledgements
- Bibliography
- About this document ...