Advertisement

Efficient Multiple Query Answering in Switched Probabilistic Relational Models

  • Marcel GehrkeEmail author
  • Tanya Braun
  • Ralf Möller
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11919)

Abstract

By accounting for context-specific independences, the size of a model can be drastically reduced, thereby making the underlying inference problem more manageable. Switched probabilistic relational models contain explicit context-specific independences. To efficiently answer multiple queries in switched probabilistic relational models, we combine the advantages of propositional gate models for context-specific independences and the lifted junction tree algorithm for answering multiple queries in probabilistic relational models. Specifically, this paper contributes (i) variable elimination in gate models, (ii) applying the lifting idea to gate models, defining switched probabilistic relational models, enabling lifted variable elimination in computations, and (iii) the switched lifted junction tree algorithm to answer multiple queries in such models efficiently. Empirical results show that using context-specific independence speeds up even lifted inference significantly.

Keywords

Lifting Context-specific independence Switched models 

References

  1. 1.
    Ahmadi, B., Kersting, K., Mladenov, M., Natarajan, S.: Exploiting symmetries for scaling loopy belief propagation and relational training. Mach. Learn. 92(1), 91–132 (2013)MathSciNetCrossRefGoogle Scholar
  2. 2.
    Boutilier, C., Friedman, N., Goldszmidt, M., Koller, D.: Context-specific independence in Bayesian networks. In: Proceedings of the 12th International Conference on Uncertainty in Artificial Intelligence, pp. 115–123. Morgan Kaufmann Publishers Inc. (1996)Google Scholar
  3. 3.
    Braun, T., Möller, R.: Lifted junction tree algorithm. In: Friedrich, G., Helmert, M., Wotawa, F. (eds.) KI 2016. LNCS (LNAI), vol. 9904, pp. 30–42. Springer, Cham (2016).  https://doi.org/10.1007/978-3-319-46073-4_3CrossRefGoogle Scholar
  4. 4.
    Cooper, G.F.: The computational complexity of probabilistic inference using Bayesian belief networks. Artif. Intell. 42(2–3), 393–405 (1990)MathSciNetCrossRefGoogle Scholar
  5. 5.
    Gogate, V., Domingos, P.M.: Probabilistic theorem proving. In: UAI 2011, Proceedings of the Twenty-Seventh Conference on Uncertainty in Artificial Intelligence, Barcelona, Spain, 14–17 July 2011, pp. 256–265. AUAI Press (2011)Google Scholar
  6. 6.
    Lauritzen, S.L., Spiegelhalter, D.J.: Local computations with probabilities on graphical structures and their application to expert systems. J. Roy. Stat. Soc.: Ser. B (Methodol.) 50(2), 157–224 (1988)MathSciNetzbMATHGoogle Scholar
  7. 7.
    Minka, T., Winn, J.: Gates. In: Advances in Neural Information Processing Systems, pp. 1073–1080 (2009)Google Scholar
  8. 8.
    Pearl, J.: Causality. Cambridge University Press, Cambridge (2009)CrossRefGoogle Scholar
  9. 9.
    Poole, D.: First-order probabilistic inference. In: Proceedings of IJCAI, vol. 3, pp. 985–991 (2003)Google Scholar
  10. 10.
    Richardson, M., Domingos, P.: Markov logic networks. Mach. Learn. 62(1), 107–136 (2006)CrossRefGoogle Scholar
  11. 11.
    Shafer, G.R., Shenoy, P.P.: Probability propagation. Ann. Math. Artif. Intell. 2(1), 327–351 (1990)MathSciNetCrossRefGoogle Scholar
  12. 12.
    Shenoy, P.P., Shafer, G.R.: Axioms for probability and belief-function propagation. Uncertain. Artif. Intell. 9, 169–198 (1990)MathSciNetGoogle Scholar
  13. 13.
    Taghipour, N., Fierens, D., Davis, J., Blockeel, H.: Lifted variable elimination: decoupling the operators from the constraint language. J. Artif. Intell. Res. 47(1), 393–439 (2013)MathSciNetCrossRefGoogle Scholar
  14. 14.
    Winn, J.: Causality with gates. In: Artificial Intelligence and Statistics, pp. 1314–1322 (2012)Google Scholar
  15. 15.
    Zhang, N.L., Poole, D.: A simple approach to Bayesian network computations. In: Proceedings of the 10th Canadian Conference on Artificial Intelligence, pp. 171–178. Springer, Heidelberg (1994)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Institute of Information SystemsUniversity of LübeckLübeckGermany

Personalised recommendations