Treewidth and the Computational Complexity of MAP Approximations

  • Johan Kwisthout
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8754)


The problem of finding the most probable explanation to a designated set of variables (the MAP problem) is a notoriously intractable problem in Bayesian networks, both to compute exactly and to approximate. It is known, both from theoretical considerations and from practical experiences, that low treewidth is typically an essential prerequisite to efficient exact computations in Bayesian networks. In this paper we investigate whether the same holds for approximating MAP. We define four notions of approximating MAP (by value, structure, rank, and expectation) and argue that all of them are intractable in general. We prove that efficient value-, structure-, and rank-approximations of MAP instances with high treewidth will violate the Exponential Time Hypothesis. In contrast, we hint that expectation-approximation can be done efficiently, even in MAP instances with high treewidth, if the most probable explanation has a high probability.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Abdelbar, A.M., Hedetniemi, S.M.: Approximating MAPs for belief networks is NP-hard and other theorems. Artificial Intelligence 102, 21–38 (1998)MathSciNetCrossRefzbMATHGoogle Scholar
  2. 2.
    Agrawal, M., Kayal, N., Saxena, N.: PRIMES is in P. Annals of Mathematics 160(2), 781–793 (2004)MathSciNetCrossRefzbMATHGoogle Scholar
  3. 3.
    Arora, S., Barak, B.: Computational Complexity: A Modern Approach. Cambridge University Press (2009)Google Scholar
  4. 4.
    Darwiche, A.: Modeling and Reasoning with Bayesian Networks. Cambridge University Press (2009)Google Scholar
  5. 5.
    De Campos, C.P.: New complexity results for MAP in Bayesian networks. In: Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence, pp. 2100–2106 (2011)Google Scholar
  6. 6.
    Downey, R.G., Fellows, M.R.: Parameterized Complexity. Springer, Berlin (1999)Google Scholar
  7. 7.
    Hamilton, M., Müller, M., van Rooij, I., Wareham, H.T.: Approximating solution structure. In: Demaine, E., Gutin, G.Z., Marx, D., Stege, U. (eds.) Structure Theory and FPT Algorithmics for Graphs, Digraphs and Hypergraphs. Dagstuhl Seminar Proceedings, vol. (07281) (2007)Google Scholar
  8. 8.
    Impagliazzo, R., Paturi, R.: On the complexity of k-SAT. Journal of Computer and System Sciences 62(2), 367–375 (2001)MathSciNetCrossRefzbMATHGoogle Scholar
  9. 9.
    Krentel, M.W.: The complexity of optimization problems. Journal of Computer and System Sciences 36, 490–509 (1988)MathSciNetCrossRefzbMATHGoogle Scholar
  10. 10.
    Kwisthout, J.: The computational complexity of probabilistic inference. Technical Report ICIS–R11003, Radboud University Nijmegen (2011)Google Scholar
  11. 11.
    Kwisthout, J.: Most probable explanations in Bayesian networks: Complexity and tractability. International Journal of Approximate Reasoning 52(9), 1452–1469 (2011)MathSciNetCrossRefzbMATHGoogle Scholar
  12. 12.
    Kwisthout, J.: Structure approximation of most probable explanations in Bayesian networks. In: van der Gaag, L.C. (ed.) ECSQARU 2013. LNCS (LNAI), vol. 7958, pp. 340–351. Springer, Heidelberg (2013)CrossRefGoogle Scholar
  13. 13.
    Kwisthout, J., Bodlaender, H.L., van der Gaag, L.C.: The necessity of bounded treewidth for efficient inference in Bayesian networks. In: Coelho, H., Studer, R., Wooldridge, M. (eds.) Proceedings of the 19th European Conference on Artificial Intelligence (ECAI 2010), pp. 237–242. IOS Press (2010)Google Scholar
  14. 14.
    Kwisthout, J.H.P., Bodlaender, H.L., van der Gaag, L.C.: The complexity of finding kth most probable explanations in probabilistic networks. In: Černá, I., Gyimóthy, T., Hromkovič, J., Jefferey, K., Králović, R., Vukolić, M., Wolf, S. (eds.) SOFSEM 2011. LNCS, vol. 6543, pp. 356–367. Springer, Heidelberg (2011)Google Scholar
  15. 15.
    Kwisthout, J., van Rooij, I.: Bridging the gap between theory and practice of approximate Bayesian inference. Cognitive Systems Research 24, 2–8 (2013)CrossRefGoogle Scholar
  16. 16.
    Marx, D.: Can you beat treewidth? In: Proceedings of the 48th Annual IEEE Symposium on Foundations of Computer Science (FOCS 2007), pp. 169–179 (2007)Google Scholar
  17. 17.
    Park, J.D., Darwiche, A.: Complexity results and approximation settings for MAP explanations. Journal of Artificial Intelligence Research 21, 101–133 (2004)MathSciNetzbMATHGoogle Scholar
  18. 18.
    Robertson, N., Seymour, P.D.: Graph minors II: Algorithmic aspects of tree-width. Journal of Algorithms 7, 309–322 (1986)MathSciNetCrossRefzbMATHGoogle Scholar
  19. 19.
    Valiant, L.G.: A theory of the learnable. Communications of the ACM 27(11), 1134–1142 (1984)CrossRefzbMATHGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Johan Kwisthout
    • 1
  1. 1.Donders Institute for Brain, Cognition and BehaviourRadboud University NijmegenNijmegenThe Netherlands

Personalised recommendations