Polynomial-time algorithms for counting and sampling Markov equivalent DAGs with applications
Article No.: 213, Pages 10140 - 10184
Abstract
Counting and sampling directed acyclic graphs from a Markov equivalence class are fundamental tasks in graphical causal analysis. In this paper we show that these tasks can be performed in polynomial time, solving a long-standing open problem in this area. Our algorithms are effective and easily implementable. As we show in experiments, these breakthroughs make thought-to-be-infeasible strategies in active learning of causal structures and causal effect identification with regard to a Markov equivalence class practically applicable.
References
[1]
Ali AhmadiTeshnizi, Saber Salehkaleybar, and Negar Kiyavash. Lazyiter: A fast algorithm for counting Markov equivalent DAGs and designing experiments. In Proceedings of the 37th International Conference on Machine Learning, ICML '20, pages 125-133, 2020.
[2]
Steen A. Andersson, David Madigan, and Michael D Perlman. A characterization of Markov equivalence classes for acyclic digraphs. The Annals of Statistics, 25(2):505-541, 1997.
[3]
Anne Berry, Richard Krueger, and Genevieve Simonet. Maximal label search algorithms to compute perfect and minimal elimination orderings. SIAM Journal on Discrete Mathematics, 23(1):428-446, 2009.
[4]
Jean RS Blair and Barry Peyton. An introduction to chordal graphs and clique trees. In Graph Theory and Sparse Matrix Computation, pages 1-29. Springer, 1993.
[5]
Graham R. Brightwell and Peter Winkler. Counting linear extensions is #P-complete. In Proceedings of the 23th Annual ACM Symposium on Theory of Computing, STOC '91, pages 175-181, 1991.
[6]
David Maxwell Chickering. Learning equivalence classes of Bayesian-network structures. Journal of Machine Learning Research, 2:445-498, 2002a.
[7]
David Maxwell Chickering. Optimal structure identification with greedy search. Journal of Machine Learning Research, 3:507-554, 2002b.
[8]
Derek G Corneil and Richard M Krueger. A unified view of graph searching. SIAM Journal on Discrete Mathematics, 22(4):1259-1276, 2008.
[9]
Gabriel A. Dirac. On rigid circuit graphs. Abhandlungen aus dem Mathematischen Seminar der Universität Hamburg, 25(1):71-76, 1961.
[10]
Frederick Eberhardt. Almost optimal intervention sets for causal discovery. In Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence, UAI'08, pages 161-168. AUAI Press, 2008.
[11]
Frederick Eberhardt, Clark Glymour, and Richard Scheines. On the number of experiments sufficient and in the worst case necessary to identify all causal relations among N variables. In Proceedings of the 21st Conference on Uncertainty in Artificial Intelligence, UAI'05, pages 178-184. AUAI Press, 2005.
[12]
Robert Ganian, Thekla Hamm, and Topi Talvitie. An efficient algorithm for counting Markov equivalent DAGs. In Proccedings of the 34th AAAI Conference on Artificial Intelligence, AAAI'20, pages 10136-10143. AAAI Press, 2020.
[13]
Robert Ganian, Thekla Hamm, and Topi Talvitie. An efficient algorithm for counting markov equivalent dags. Artificial Intelligence, 304:103648, 2022.
[14]
AmirEmad Ghassami, Saber Salehkaleybar, Negar Kiyavash, and Elias Bareinboim. Budgeted experiment design for causal structure learning. In Proceedings of the 35th International Conference on Machine Learning, ICML '18, pages 1719-1728, 2018.
[15]
AmirEmad Ghassami, Saber Salehkaleybar, Negar Kiyavash, and Kun Zhang. Counting and sampling from Markov equivalent DAGs using clique trees. In Proccedings of the 33th AAAI Conference on Artificial Intelligence, AAAI'19, pages 3664-3671. AAAI Press, 2019.
[16]
Steven B. Gillispie and Michael D. Perlman. The size distribution for Markov equivalence classes of acyclic digraph models. Artificial Intelligence, 141(1/2):137-155, 2002.
[17]
Kristjan H. Greenewald, Dmitriy Katz, Karthikeyan Shanmugam, Sara Magliacane, Murat Kocaoglu, Enric Boix Adserà, and Guy Bresler. Sample efficient active learning of causal trees. In Proceedings of the 32nd Conference on Neural Information Processing Systems, NeurIPS'19, pages 14279-14289, 2019.
[18]
Alain Hauser and Peter Bühlmann. Characterization and greedy learning of interventional Markov equivalence classes of directed acyclic graphs. Journal of Machine Learning Research, 13:2409-2464, 2012.
[19]
Alain Hauser and Peter Bühlmann. Two optimal strategies for active learning of causal models from interventional data. International Journal of Approximate Reasoning, 55 (4):926-939, 2014.
[20]
Yang-Bo He and Zhi Geng. Active learning of causal networks with intervention experiments and optimal designs. Journal of Machine Learning Research, 9(Nov):2523-2547, 2008.
[21]
Yangbo He and Bin Yu. Formulas for counting the sizes of Markov equivalence classes of directed acyclic graphs. arXiv, abs/1610.07921, 2016. URL http://arxiv.org/abs/1610.07921.
[22]
Yangbo He, Jinzhu Jia, and Bin Yu. Counting and exploring sizes of Markov equivalence classes of directed acyclic graphs. Journal of Machine Learning Research, 16(79):2589- 2609, 2015.
[23]
David Heckerman, Dan Geiger, and David Maxwell Chickering. Learning Bayesian networks: The combination of knowledge and statistical data. Machine Learning, 20(3):197-243, 1995.
[24]
Arthur B Kahn. Topological sorting of large networks. Communications of the ACM, 5 (11):558-562, 1962.
[25]
Daphne Koller and Nir Friedman. Probabilistic Graphical Models - Principles and Techniques. MIT Press, 2009. ISBN 978-0-262-01319-2.
[26]
Marloes H Maathuis, Markus Kalisch, and Peter Bühlmann. Estimating high-dimensional intervention effects from observational data. The Annals of Statistics, 37(6A):3133-3164, 2009.
[27]
Christopher Meek. Causal inference and causal explanation with background knowledge. In Proceedings of the 11th Conference on Uncertainty in Artificial Intelligence, UAI'95, pages 403-410, 1995.
[28]
Christopher Meek. Graphical Models: Selecting Causal and Statistical Models. PhD thesis, Carnegie Mellon University, 1997.
[29]
OEIS Foundation Inc. The number of irreducible permutations. entry a003319 in the on-line encyclopedia of integer sequences, 2022. URL https://oeis.org/A003319.
[30]
Judea Pearl. Causality. Cambridge University Press, 2009. ISBN 978-0521895606.
[31]
Emilija Perković, Johannes Textor, Markus Kalisch, and Marloes H. Maathuis. Complete graphical characterization and construction of adjustment sets in Markov equivalence classes of ancestral graphs. Journal of Machine Learning Research, 18:220:1-220:62, 2017.
[32]
Donald J. Rose, Robert Endre Tarjan, and George S. Lueker. Algorithmic aspects of vertex elimination on graphs. SIAM Journal on Computing, 5(2):266-283, 1976.
[33]
Edward R. Scheinerman. Random interval graphs. Combinatorica, 8(4):357-371, 1988.
[34]
Oylum Seker, Pinar Heggernes, Tínaz Ekim, and Z. Caner Taskin. Linear-time generation of random chordal graphs. In Proccedings of the 10th International Conference on Algorithms and Complexity, CIAC 17, volume 10236, pages 442-453, 2017.
[35]
Karthikeyan Shanmugam, Murat Kocaoglu, Alexandros G. Dimakis, and Sriram Vishwanath. Learning causal graphs with small interventions. In Processing of the 28th Conference on Neural Information Processing Systems, NeurIPS'15, pages 3195-3203, 2015.
[36]
Ilya Shpitser and Judea Pearl. Identification of joint interventional distributions in recursive semi-Markovian causal models. In Proceedings of the 21st AAAI Conference on Artificial Intelligence, AAAI'06, volume 21, pages 1219-1226. AAAI Press, 2006.
[37]
Ilya Shpitser, Tyler VanderWeele, and James Robins. On the validity of covariate adjustment for estimating causal effects. In Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, UAI'10, pages 527-536. AUAI Press, 2010.
[38]
Peter Spirtes, Clark Glymour, and Richard Scheines. Causation, Prediction, and Search, Second Edition. MIT Press, 2000. ISBN 978-0-262-19440-2.
[39]
Chandler Squires, Sara Magliacane, Kristjan Greenewald, Dmitriy Katz, Murat Kocaoglu, and Karthikeyan Shanmugam. Active structure learning of causal DAGs via directed clique trees. In Proceedings of the 33rd Conference on Neural Information Pressing Systems, NeurIPS'20, volume 33, pages 21500-21511, 2020.
[40]
Topi Talvitie and Mikko Koivisto. Counting and sampling Markov equivalent directed acyclic graphs. In Proccedings of the 33th AAAI Conference on Artificial Intelligence, AAAI'19, pages 7984-7991. AAAI Press, 2019.
[41]
Robert E Tarjan and Mihalis Yannakakis. Simple linear-time algorithms to test chordality of graphs, test acyclicity of hypergraphs, and selectively reduce acyclic hypergraphs. SIAM Journal on computing, 13(3):566-579, 1984.
[42]
Benito van der Zander and Maciej Liśkiewicz. Separators and adjustment sets in Markov equivalent DAGs. In Proceedings of the 30th AAAI Conference on Artificial Intelligence, AAAI'16, pages 3315-3321. AAAI Press, 2016.
[43]
Benito van der Zander, Maciej Liśkiewicz, and Johannes Textor. Separators and adjustment sets in causal graphs: Complete criteria and an algorithmic framework. Artificial Intelligence, 270:1-40, 2019.
[44]
Thomas Verma and Judea Pearl. Equivalence and synthesis of causal models. In Proceedings of the 6th Conference on Uncertainty in Artificial Intelligence, UAI'90, pages 255-270, 1990.
[45]
Thomas Verma and Judea Pearl. An algorithm for deciding if a set of observed independencies has a causal explanation. In Proceedings of the 8th Conference on Uncertainty in Artificial Intelligence, UAI'92, pages 323-330, 1992.
[46]
Michael D Vose. A linear algorithm for generating random numbers with a given distribution. IEEE Transactions on software engineering, 17(9):972-975, 1991.
[47]
Alastair J Walker. New fast method for generating discrete random numbers with arbitrary frequency distributions. Electronics Letters, 10(8):127-128, 1974.
[48]
Marcel Wienöbst, Max Bannach, and Maciej Liśkiewicz. Extendability of causal graphical models: Algorithms and computational complexity. In Proceedings of the 37th Conference on Uncertainty in Artificial Intelligence, UAI'21. AUAI Press, 2021a.
[49]
Marcel Wienöbst, Max Bannach, and Maciej Liśkiewicz. Polynomial-time algorithms for counting and sampling Markov equivalent DAGs. In Proceedings of the 35th AAAI Conference on Artificial Intelligence, AAAI'21, pages 12198-12206. AAAI Press, 2021b.
Index Terms
- Polynomial-time algorithms for counting and sampling Markov equivalent DAGs with applications
Index terms have been assigned to the content through auto-classification.
Recommendations
Characterization and greedy learning of interventional Markov equivalence classes of directed acyclic graphs
The investigation of directed acyclic graphs (DAGs) encoding the same Markov property, that is the same conditional independence relations of multivariate observational distributions, has a long tradition; many algorithms exist for model selection and ...
Counting and sampling from Markov equivalent DAGs using clique trees
AAAI'19/IAAI'19/EAAI'19: Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence and Thirty-First Innovative Applications of Artificial Intelligence Conference and Ninth AAAI Symposium on Educational Advances in Artificial IntelligenceA directed acyclic graph (DAG) is the most common graphical model for representing causal relationships among a set of variables. When restricted to using only observational data, the structure of the ground truth DAG is identifiable only up to Markov ...
Efficient Domination for Some Subclasses of $$P_6$$-free Graphs in Polynomial Time
WG 2015: Revised Papers of the 41st International Workshop on Graph-Theoretic Concepts in Computer Science - Volume 9224Let G be a finite undirected graph. A vertex dominates itself and all its neighbors in G. A vertex set D is an efficient dominating sete.d. for short of G if every vertex of G is dominated by exactly one vertex of D. The Efficient Domination ED problem, ...
Comments
Information & Contributors
Information
Published In
January 2023
18881 pages
Copyright © 2023.
CC-BY 4.0
Publisher
JMLR.org
Publication History
Published: 06 March 2024
Accepted: 01 July 2023
Revised: 01 April 2023
Received: 01 May 2022
Published in JMLR Volume 24, Issue 1
Author Tags
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 24Total Downloads
- Downloads (Last 12 months)24
- Downloads (Last 6 weeks)13
Reflects downloads up to 04 Oct 2024
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in