Creating non-minimal triangulations for use in inference in mixed stochastic/deterministic graphical models

被引:0
作者
Chris D. Bartels
Jeff A. Bilmes
机构
[1] University of Washington,Department of Electrical Engineering
来源
Machine Learning | 2011年 / 84卷
关键词
Triangulation; Inference; Graphical models; Graph theory;
D O I
暂无
中图分类号
学科分类号
摘要
We demonstrate that certain large-clique graph triangulations can be useful for reducing computational requirements when making queries on mixed stochastic/deterministic graphical models. This is counter to the conventional wisdom that triangulations that minimize clique size are always most desirable for use in computing queries on graphical models. Many of these large-clique triangulations are non-minimal and are thus unattainable via the popular elimination algorithm. We introduce ancestral pairs as the basis for novel triangulation heuristics and prove that no more than the addition of edges between ancestral pairs needs to be considered when searching for state space optimal triangulations in such graphs. Empirical results on random and real world graphs are given. We also present an algorithm and correctness proof for determining if a triangulation can be obtained via elimination, and we show that the decision problem associated with finding optimal state space triangulations in this mixed setting is NP-complete.
引用
收藏
页码:249 / 289
页数:40
相关论文
共 47 条
[11]  
Davis M.(1972)Algorithms for minimum coloring, maximum clique, minimum covering by cliques, and maximum independent set of a chordal graph SIAM Journal on Computing 1 180-187
[12]  
Logemann G.(1990)Bayesian updating in causal probabilistic networks by local computation CSQ. Computational Statistics Quarterly 4 269-282
[13]  
Loveland D.(1997)Decomposing Bayesian networks by genetic algorithms Statistics and Computing 7 19-34
[14]  
Dean T.(1988)Local computations with probabilities on graphical structures and their application to expert systems The Journal of the Royal Statistical Society 50 57-224
[15]  
Kanazawa K.(1998)An efficient algorithm for finding the m most probable configurations in probabilistic expert systems Statistics and Computing 8 159-173
[16]  
Dechter R.(1976)Minimal triangulation of a graph and optimal pivoting order in a sparse matrix Journal of Mathematical Analysis and Applications 54 622-633
[17]  
Dechter R.(2002)Maximal prime subgraph decomposition of Bayesian networks IEEE Transactions on Systems, Man and Cybernetics. Part B. Cybernetics 32 21-31
[18]  
Fattah Y.E.(1961)The use of linear graphs in gauss elimination SIAM Review 3 119-130
[19]  
Fishelson M.(1970)Triangulated graphs and the elimination process Journal of Mathematical Analysis and Applications 32 597-609
[20]  
Geiger D.(1976)Algorithmic aspects of vertex elimination on graphs SIAM Journal on Computing 5 266-283