Abstract
It is well-known that the observation of a variable in a Bay-esian network can affect the effective connectivity of the network, which in turn affects the efficiency of inference. Unfortunately, the observed variables may not be known until runtime, which limits the amount of compile-time optimization that can be done in this regard. In this paper, we consider how to improve inference when we know the likelihood of a variable being observed. We show how these probabilities of observation can be exploited to improve existing heuristics for choosing elimination orderings for inference. Empirical tests over a set of benchmark networks using the Variable Elimination algorithm show reductions of up to 50%, 70%, and 55% in multiplications, summations, and runtime, respectively.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Becker, A., Geiger, D.: A Sufficiently Fast Algorithm for Finding Close to Optimal Clique Trees. Artificial Intelligence 125, 3–17 (2001)
Cooper, G.F.: The Computational Complexity of Probabilistic Inference using Bayesian Belief Networks. Artificial Intelligence 42, 393–405 (1990)
Darwiche, A.: Recursive Conditioning - Any-space Conditioning Method with Treewidth-Bounded Complexity. Artificial Intelligence 126(1-2), 5–41 (2000)
Dow, P.A., Korf, R.E.: Best-First Search for Treewidth. In: Proceedings of the 22nd Conference on Aritificial Intelligence. AAAI press, CA (2007)
Fishelson, M., Geiger, D.: Optimizing Exact Genetic Linkage Computations. In: Proceedings of the International Conference on Research in Computational Molecular Biology (RECOMB), pp. 114–121 (April 2003)
Gogate, V., Dechter, R.: A Complete Anytime Algorithm for Treewidth. In: Proceedings of the 20th Annual Conference on Uncertainity in Artificial Intelligence, pp. 201–208. AUAI press, Arlinton (2004)
Kjaerulff, U.: Triangulation of Graphs—Algorithms Giving Small Total State Space. Tech. Rep. R90-09, Department of Computer Science, University of Aalborg, Denmark (March 1990)
Kjaerulff, U.B., Madsen, A.L.: Bayesian Networks and Influence Diagrams: A Guide to Construction and Analysis. Springer Publishing Company (2007)
Koller, D., Friedman, N.: Probabilistic Graphical Models - Principles and Techniques. MIT Press (2009)
Langevin, S., Valtorta, M.: Performance Evaluation of Algorithms for Soft Evidential Update in Bayesian Networks: First Results. In: Greco, S., Lukasiewicz, T. (eds.) SUM 2008. LNCS (LNAI), vol. 5291, pp. 284–297. Springer, Heidelberg (2008)
Lauritzen, S., Spiegelhalter, D.: Local Computations with Probabilities on Graphical Structures and Their Application to Expert Systems. Journal of the Royal Statistical Society, Series B 50, 157–224 (1988)
Pan, R., Peng, Y., Ding, Z.: Belief Update in Bayesian Networks Using Uncertain Evidence. In: 18th IEEE International Conference on Tools with Artificial Intelligence, pp. 441–444 (2006)
Shoikhet, K., Geiger, D.: A Practical Algorithm for Finding Optimal Triangulations. In: Proceedings of the AAAI/IAAI, pp. 185–190 (1997)
Zhang, N., Poole, D.: A Simple Approach to Bayesian Network Computations. In: The 10th Canadian Conference on Artificial Intelligence, pp. 171–178 (1994)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Mousumi, F., Grant, K. (2012). Exploiting the Probability of Observation for Efficient Bayesian Network Inference. In: Kosseim, L., Inkpen, D. (eds) Advances in Artificial Intelligence. Canadian AI 2012. Lecture Notes in Computer Science(), vol 7310. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-30353-1_12
Download citation
DOI: https://doi.org/10.1007/978-3-642-30353-1_12
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-30352-4
Online ISBN: 978-3-642-30353-1
eBook Packages: Computer ScienceComputer Science (R0)