Abstract
PDF (157 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (3)
Purchase PDF (69 K)
doi:10.1016/S0004-3702(02)00360-0 
Copyright © 2002 Elsevier Science B.V. All rights reserved.
Maximum likelihood bounded tree-width Markov networks*1
Nathan Srebro
Received 17 December 2001.
Abstract
We study the problem of projecting a distribution onto (or finding a maximum likelihood distribution among) Markov networks of bounded tree-width. By casting it as the combinatorial optimization problem of finding a maximum weight hypertree, we prove that it is NP-hard to solve exactly and provide an approximation algorithm with a provable performance guarantee.
Author Keywords: Markov networks; Markov random fields; Undirected graphical models; Entropy decomposition; Hyper-trees; Tree-width; Hardness
References
1. S. Amari , Information geometry on hierarchy of probability distributions. IEEE Trans. Inform. Theory 47 5 (2001), pp. 1701–1711. View Record in Scopus | Cited By in Scopus (50)
2. N. Alon and J.H. Spencer , The Probabilistic Method. , Wiley, New York (1991).
3. J. Besag , Spatial interaction and the statistical analysis of lattice systems. Proc. Roy. Statist. Soc. Ser. B (1974), pp. 192–236.
4. H.L. Bodlaender , A linear time algorithm for finding tree-decompositions of small treewidth. SIAM J. Comput. 25 (1996), pp. 1305–1317. View Record in Scopus | Cited By in Scopus (210)
5. D.M. Chickering , Learning Bayesian networks is NP-complete. In: D. Fisher and H.-J. Lenz, Editors, Learning from Data: AI and Statistics, Vol. V, Springer-Verlag, New York (1996), pp. 121–130.
6. C.K. Chow and C.N. Liu , Approximating discrete probability distributions with dependence trees. IEEE Trans. Inform. Theory IT-14 3 (1968), pp. 462–467.
7. T.M. Cover and J.A. Thomas , Elements of Information Theory. , Wiley–Intersciences, New York (1991).
8. S. Dasgupta , Learning polytrees. In: Proc. 15th Conf. on Uncertainty in Artificial Intelligence (UAI-99), Stockholm, Sweden (1999), pp. 134–141.
9. K.U. Höffgen , Learning and robust learning of product distributions. In: Proceedings of the Sixth Annual Workshop on Computational Learning Theory, Santa Cruz, CA (1993), pp. 77–83.
10. D. Karger and N. Srebro , Learning Markov networks: Maximum bounded tree-width graphs. In: Proceedings of the 12th ACM–SIAM Symposium on Discrete Algorithms (2001).
11. F.M. Malvestuto , Approximating discrete probability distributions with decomposable models. IEEE Trans. Syst. Man Cybernetics 21 5 (1991), pp. 1287–1294. View Record in Scopus | Cited By in Scopus (10)
12. M. Meila-Predoviciu, Learning with mixtures of trees, Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA, 1999.
13. J. Pearl , Probabilistic Reasoning in Intelligent Systems. (Revised 2nd Printing Edition ed.),, Morgan Kaufmann, San Mateo, CA (1997).
14. K. Shoikhet and D. Geiger , A practical algorithm for finding optimal triangulations. In: Proceedings AAAI-97, Providence, RI (1997), pp. 185–190. View Record in Scopus | Cited By in Scopus (16)
15. N. Srebro, Maximum likelihood Markov networks: An algorithmic approach, Master's Thesis, Massachusetts Institute of Technology, Cambridge, MA, 2000.
16. N. Wermuth and S. Lauritzen , Graphical and recursive models of contingency tables. Biometrika 72 (1983), pp. 537–552.
*1 This is an extended version of the paper presented at the 17th Conference on Uncertainty in Artificial Intelligence (UAI-01), Seattle, WA, 2001.
