Decision-theoretic specification of credal networks: A unified language for uncertain modeling with sets of Bayesian networks

Credal networks are models that extend Bayesian nets to deal with imprecision in probability, and can actually be regarded as sets of Bayesian nets. Credal nets appear to be powerful means to represent and deal with many important and challenging problems in uncertain reasoning. We give examples to show that some of these problems can only be modeled by credal nets called non-separately specified. These, however, are still missing a graphical representation language and updating algorithms. The situation is quite the opposite with separately specified credal nets, which have been the subject of much study and algorithmic development. This paper gives two major contributions. First, it delivers a new graphical language to formulate any type of credal network, both separately and non-separately specified. Second, it shows that any non-separately specified net represented with the new language can be easily transformed into an equivalent separately specified net, defined over a larger domain. This result opens up a number of new outlooks and concrete outcomes: first of all, it immediately enables the existing algorithms for separately specified credal nets to be applied to non-separately specified ones. We explore this possibility for the 2U algorithm: an algorithm for exact updating of singly connected credal nets, which is extended by our results to a class of non-separately specified models. We also consider the problem of inference on Bayesian networks, when the reason that prevents some of the variables from being observed is unknown. The problem is first reformulated in the new graphical language, and then mapped into an equivalent problem on a separately specified net. This provides a first algorithmic approach to this kind of inference, which is also proved to be NP-hard by similar transformations based on our formalism.

[1]  Marco Zaffalon,et al.  Equivalence Between Bayesian and Credal Nets on an Updating Problem , 2006, SMPS.

[2]  P. Walley.  Statistical Reasoning with Imprecise Probabilities , 1990 .

[3]  Gert de Cooman,et al.  Updating beliefs with incomplete observations , 2003, Artif. Intell..

[4]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[5]  Marco Zaffalon,et al.  Locally specified credal networks , 2006, Probabilistic Graphical Models.

[6]  Isaac Levi,et al.  The Enterprise Of Knowledge , 1980 .

[7]  Andrés Cano,et al.  Credal Nets with Probabilities Estimated with an Extreme Imprecise Dirichlet Model , 2007 .

[8]  Judea Pearl,et al.  Equivalence and Synthesis of Causal Models , 1990, UAI.

[9]  Serafín Moral,et al.  Using probability trees to compute marginals with imprecise probabilities , 2002, Int. J. Approx. Reason..

[10]  Fabio Gagliardi Cozman,et al.  The Inferential Complexity of Bayesian and Credal Networks , 2005, IJCAI.

[11]  Fabio Gagliardi Cozman,et al.  Graphical models for imprecise probabilities , 2005, Int. J. Approx. Reason..

[12]  Fabio Gagliardi Cozman,et al.  Inference with Seperately Specified Sets of Probabilities in Credal Networks , 2002, UAI.

[13]  Fabio Gagliardi Cozman,et al.  Credal networks , 2000, Artif. Intell..

[14]  Enrico Fagiuoli,et al.  2U: An Exact Interval Propagation Algorithm for Polytrees with Binary Variables , 1998, Artif. Intell..

[15]  Fabio Gagliardi Cozman,et al.  Propositional and Relational Bayesian Networks Associated with Imprecise and Qualitat , 2004, UAI.

[16]  Andrés Cano,et al.  Convex Sets Of Probabilities Propagation By Simulated Annealing , 1994 .

[17]  Marco Zaffalon,et al.  Fast algorithms for robust classification with Bayesian nets , 2007, Int. J. Approx. Reason..

[18]  P. Walley.  Inferences from Multinomial Data: Learning About a Bag of Marbles , 1996 .

[19]  Marco Zaffalon,et al.  Conservative Rules for Predictive Inference with Incomplete Data , 2005, ISIPTA.

[20]  Marco Zaffalon,et al.  Statistical inference of the naive credal classifier , 2001, ISIPTA.

[21]  Serafín Moral,et al.  Strong Conditional Independence for Credal Sets , 2004, Annals of Mathematics and Artificial Intelligence.

[22]  D. Rubin,et al.  Statistical Analysis with Missing Data , 1988 .

[23]  Michael P. Wellman.  Fundamental Concepts of Qualitative Probabilistic Networks , 1990, Artif. Intell..

[24]  Nevin Lianwen Zhang,et al.  A computational theory of decision networks , 1993, Int. J. Approx. Reason..

[25]  David S. Johnson,et al.  Computers and Intractability: A Guide to the Theory of NP-Completeness , 1978 .

[26]  Fabio Gagliardi Cozman,et al.  Binarization Algorithms for Approximate Updating in Credal Nets , 2006, STAIRS.