{"type":"ScholarlyArticle","title":"A Core-Guided Approach to Learning Optimal Causal Graphs","authors":[{"name":"Antti Hyttinen","slug":"antti-hyttinen","given":"Antti","family":"Hyttinen","affiliation":"Department of Computer Science , University of Helsinki , Finland"},{"name":"Paul Saikko","slug":"paul-saikko","given":"Paul","family":"Saikko","affiliation":"Department of Computer Science , University of Helsinki , Finland"},{"name":"Matti Järvisalo","slug":"matti-jarvisalo","given":"Matti","family":"Järvisalo","affiliation":"Department of Computer Science , University of Helsinki , Finland"}],"abstract":"Discovery of causal relations is an important part of data analysis. Recent exact Boolean optimization approaches enable tackling very general search spaces of causal graphs with feedback cycles and latent confounders, simultaneously obtaining high accuracy by optimally combining conflicting independence information in sample data. We propose several domain-specific techniques and integrate them into a core-guided maximum satisfiability solver, thereby speeding up current state of the art in exact search for causal graphs with cycles and latent confounders on simulated and real-world data.","doi":"10.24963/ijcai.2017/90","url":"https://ijcai.org/proceedings/2017/90","pdf":"https://ijcai.org/proceedings/2017/0090.pdf","markdown":"https://ijcai.org/proceedings/2017/90.md","bibtex":"https://ijcai.org/proceedings/2017/bibtex/90","volume":{"slug":"2017","title":"Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2017","url":"https://ijcai.org/proceedings/2017/","year":2017,"location":"Melbourne, Australia","dates":"19-25 August 2017","editors":"Carles Sierra","isbn":"978-0-9992411-0-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2017"},"section":"Constraints and Satisfiability","track":"Main track","number":90,"pages":{"first":645,"last":651},"keywords":["Knowledge Representation, Reasoning, and Logic: Action, Change and Causality","Constraints and Satisfiability: MaxSAT, MinSAT","Machine Learning: Learning Graphical Models","Constraints and Satisfiability: Constraint Optimisation"],"videos":[],"references":[{"position":1,"raw":"[Alviano et al., 2015] M. Alviano, C. Dodaro, and F. Ricca. A MaxSAT algorithm using cardinality constraints of bounded size. In IJCAI. AAAI Press, 2015.","title":"A MaxSAT algorithm using cardinality constraints of bounded size","authors":"M Alviano, C Dodaro, F Ricca","year":"2015","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2015/379"},{"position":2,"raw":"[Ansótegui and Gabàs, 2013] C. Ansótegui and J. Gabàs. Solving (weighted) partial MaxSAT with ILP. In CPAIOR, volume 7874 of LNCS, pages 403-409. Springer, 2013.","title":"Solving (weighted) partial MaxSAT with ILP","authors":"C Ansótegui, J Gabàs","year":"2013","venue":"CPAIOR","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Ansótegui et al., 2013] C. Ansótegui, M.L. Bonet, J. Gabàs, and J. Levy. Improving WPM2 for (weighted) partial MaxSAT. In CP, volume 8124 of LNCS, pages 117-132. Springer, 2013.","title":"Improving WPM2 for (weighted) partial MaxSAT","authors":"C Ansótegui, M L Bonet, J Gabàs, J Levy","year":"2013","venue":"CP","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Ansótegui et al., 2016] C. Ansótegui, J. Gabàs, and J. Levy. Exploiting subproblem optimization in SAT-based MaxSAT algorithms. J. Heuristics, 22(1):1-53, 2016.","title":"Exploiting subproblem optimization in SAT-based MaxSAT algorithms","authors":"C Ansótegui, J Gabàs, J Levy","year":"2016","venue":"J. Heuristics","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Bartlett and Cussens, 2017] M. Bartlett and J. Cussens. In- teger linear programming for the Bayesian network struc- ture learning problem. Artif. Intell., 244:258-271, 2017.","title":"Integer linear programming for the Bayesian network structure learning problem","authors":"M Bartlett, J Cussens","year":"2017","venue":"Artif. Intell","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Berg et al., 2014] J. Berg, M. Järvisalo, and B. Malone. Learning optimal bounded treewidth Bayesian networks via maximum satisfiability. In AISTATS, volume 33 of JMLR W&CP, pages 86-95. JMLR, 2014.","title":"Learning optimal bounded treewidth Bayesian networks via maximum satisfiability","authors":"J Berg, M Järvisalo, B Malone","year":"2014","venue":"JMLR W&CP","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Berg et al., 2015] J. Berg, A. Hyttinen, and M. Järvisalo. Applications of MaxSAT in data analysis. In PoS, 2015.","title":"Applications of MaxSAT in data analysis","authors":"J Berg, A Hyttinen, M Järvisalo","year":"2015","venue":"PoS","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Borboudakis and Tsamardinos, 2016] G. Borboudakis and I. Tsamardinos. Towards robust and versatile causal dis- covery for business applications. In KDD. ACM, 2016.","title":"Towards robust and versatile causal discovery for business applications","authors":"G Borboudakis, I Tsamardinos","year":"2016","venue":"KDD. ACM","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Claassen and Heskes, 2012] T. Claassen and T. Heskes. A Bayesian approach to constraint based causal inference. In UAI, pages 207-216. AUAI Press, 2012.","title":"A Bayesian approach to constraint based causal inference","authors":"T Claassen, T Heskes","year":"2012","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Crowder et al., 1983] H. Crowder, E.L. Johnson, and M. Padberg. Solving large-scale zero-one linear program- ming problems. Oper. Res., 31(5):803-834, 1983.","title":"Solving large-scale zero-one linear programming problems","authors":"H Crowder, E L Johnson, M Padberg","year":"1983","venue":"Oper. Res","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Danzig et al., 1954] G.B. Danzig, D.R. Fulkerson, and S.M. Johnson. Solution of a large-scale traveling-salesman problem. Oper. Res., 2:393-410, 1954.","title":"Solution of a large-scale traveling-salesman problem","authors":"G B Danzig, D R Fulkerson, S M Johnson","year":"1954","venue":"Oper. Res","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Davies and Bacchus, 2011] J. Davies and F. Bacchus. Solv- ing MAXSAT by solving a sequence of simpler SAT in- stances. In CP, volume 6876 of LNCS, pages 225-239. Springer, 2011.","title":"Solving MAXSAT by solving a sequence of simpler SAT instances","authors":"J Davies, F Bacchus","year":"2011","venue":"CP","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Davies and Bacchus, 2013] J. Davies and F. Bacchus. Ex- ploiting the power of MIP solvers in MAXSAT. In SAT, volume 7962 of LNCS, pages 166-181. Springer, 2013.","title":"Exploiting the power of MIP solvers in MAXSAT","authors":"J Davies, F Bacchus","year":"2013","venue":"SAT","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Eén and Sörensson, 2004] N. Eén and N. Sörensson. An ex- tensible SAT-solver. In SAT 2003, volume 2919 of LNCS, pages 502-518. Springer, 2004.","title":"An extensible SAT-solver","authors":"N Eén, N Sörensson","year":"2004","venue":"SAT 2003","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Gebser et al., 2012] M. Gebser, B. Kaufmann, and T. Schaub. Conflict-driven answer set solving: From theory to practice. Artif. Intell., 187:52-89, 2012.","title":"Conflict-driven answer set solving: From theory to practice","authors":"M Gebser, B Kaufmann, T Schaub","year":"2012","venue":"Artif. Intell","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Hyttinen et al., 2013] A. Hyttinen, P. O. Hoyer, F. Eber- hardt, and M. Järvisalo. Discovering cyclic causal models with latent variables: A general SAT-based procedure. In UAI, pages 301-310. AUAI Press, 2013.","title":"Discovering cyclic causal models with latent variables: A general SAT-based procedure","authors":"A Hyttinen, P O Hoyer, F Eberhardt, M Järvisalo","year":"2013","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Hyttinen et al., 2014] A. Hyttinen, F. Eberhardt, and M. Järvisalo. Constraint-based causal discovery: Conflict resolution with answer set programming. In UAI, pages 340-349. AUAI Press, 2014.","title":"Constraint-based causal discovery: Conflict resolution with answer set programming","authors":"A Hyttinen, F Eberhardt, M Järvisalo","year":"2014","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Koshimura et al., 2012] M. Koshimura, T. Zhang, H. Fujita, and R. Hasegawa. QMaxSAT: A partial Max-SAT solver. JSAT, 8(1/2):95-100, 2012.","title":"QMaxSAT: A partial Max-SAT solver","authors":"M Koshimura, T Zhang, H Fujita, R Hasegawa","year":"2012","venue":"JSAT","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Lagniez and Biere, 2013] J.-M. Lagniez and A. Biere. Fac- toring out assumptions to speed up MUS extraction. In SAT, volume 7962 of LNCS, pages 276-292. Springer, 2013.","title":"Factoring out assumptions to speed up MUS extraction","authors":"J.-M Lagniez, A Biere","year":"2013","venue":"SAT","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Magliacane et al., 2016] S. Magliacane, T. Claassen, and J.M. Mooij. Ancestral causal inference. In NIPS, pages 4466-4474, 2016.","title":"Ancestral causal inference","authors":"S Magliacane, T Claassen, J M Mooij","year":"2016","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Malone et al., 2015] B. Malone, M. Järvisalo, and P. Myl- lymäki. Impact of learning strategies on the quality of Bayesian networks: An empirical evaluation. In UAI, pages 562-571. AUAI Press, 2015.","title":"Impact of learning strategies on the quality of Bayesian networks: An empirical evaluation","authors":"B Malone, M Järvisalo, P","year":"2015","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Margaritis and Bromberg, 2009] D. Margaritis and F. Bromberg. Efficient Markov network discovery using particle filters. Comput. Intell., 25(4):367-394, 2009.","title":"Efficient Markov network discovery using particle filters","authors":"D Margaritis, F Bromberg","year":"2009","venue":"Comput. Intell","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Martins et al., 2014] R. Martins, V.M. Manquinho, and I. Lynce. Open-WBO: A modular MaxSAT solver. In SAT, volume 8561 of LNCS, pages 438-445. Springer, 2014.","title":"Open-WBO: A modular MaxSAT solver","authors":"R Martins, V M Manquinho, I Lynce","year":"2014","venue":"SAT","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Moreno-Centeno and Karp, 2013] E. Moreno-Centeno and R.M. Karp. The implicit hitting set approach to solve combinatorial optimization problems with an application to multigenome alignment. Oper. Res., 61(2):453-468, 2013.","title":"The implicit hitting set approach to solve combinatorial optimization problems with an application to multigenome alignment","authors":"Moreno-Centeno, ] E Karp, R M Moreno-Centeno, Karp","year":"2013","venue":"Oper. Res","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Morgado et al., 2015] A. Morgado, A. Ignatiev, and J. Marques-Silva. MSCG: Robust core-guided MaxSAT solving. JSAT, 9:129-134, 2015.","title":"MSCG: Robust core-guided MaxSAT solving","authors":"A Morgado, A Ignatiev, J Marques-Silva","year":"2015","venue":"JSAT","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Neal, 2000] R. Neal. On deducing conditional indepen- dence from d-separation in causal graphs with feedback. J. Artif. Intell. Res., 12:87-91, 2000.","title":"On deducing conditional independence from d-separation in causal graphs with feedback","authors":"R Neal","year":"2000","venue":"J. Artif. Intell. Res","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Pearl and Dechter, 1996] J. Pearl and R. Dechter. Identi- fying independencies in causal graphs with feedback. In UAI, pages 420-426.","title":"Identifying independencies in causal graphs with feedback","authors":"J Pearl, R Dechter","year":null,"venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"Morgan Kaufmann, 1996.","title":null,"authors":"Morgan Kaufmann","year":"1996","venue":null,"doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Pearl, 2000] J. Pearl. Causality: Models, Reasoning, and Inference. Cambridge University Press, 2000.","title":"Causality: Models, Reasoning, and Inference","authors":"J Pearl","year":"2000","venue":null,"doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Richardson, 1996] T. Richardson. A discovery algorithm for directed cyclic graphs. In UAI, pages 454-461.","title":"A discovery algorithm for directed cyclic graphs","authors":"T Richardson","year":null,"venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":31,"raw":"Morgan Kaufmann, 1996.","title":null,"authors":"Morgan Kaufmann","year":"1996","venue":null,"doi":null,"url":null,"internal_url":null},{"position":32,"raw":"[Saikko et al., 2016] P. Saikko, J. Berg, and M. Järvisalo. LMHS: A SAT-IP hybrid MaxSAT solver. In SAT, volume 9710 of LNCS, pages 539-546. Springer, 2016.","title":"LMHS: A SAT-IP hybrid MaxSAT solver","authors":"P Saikko, J Berg, M Järvisalo","year":"2016","venue":"SAT","doi":null,"url":null,"internal_url":null},{"position":33,"raw":"[Spirtes et al., 2000] P. Spirtes, C. Glymour, and R. Scheines. Causation, Prediction, and Search. MIT Press, 2000.","title":"Causation, Prediction, and Search","authors":"P Spirtes, C Glymour, R Scheines","year":"2000","venue":null,"doi":null,"url":null,"internal_url":null},{"position":34,"raw":"[Spirtes, 1995] P. Spirtes. Directed cyclic graphical repre- sentation of feedback models. In UAI, pages 491-498.","title":"Directed cyclic graphical representation of feedback models","authors":"P Spirtes","year":null,"venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":35,"raw":"Morgan Kaufmann, 1995.","title":null,"authors":"Morgan Kaufmann","year":"1995","venue":null,"doi":null,"url":null,"internal_url":null},{"position":36,"raw":"[Studený, 1998] M. Studený. Bayesian networks from the point of view of chain graphs. In UAI, pages 496-503.","title":"Bayesian networks from the point of view of chain graphs","authors":"M Studený","year":null,"venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":37,"raw":"Morgan Kaufmann, 1998.","title":null,"authors":"Morgan Kaufmann","year":"1998","venue":null,"doi":null,"url":null,"internal_url":null},{"position":38,"raw":"[Triantafillou and Tsamardinos, 2015] S. Triantafillou and I. Tsamardinos. Constraint-based causal discovery from multiple interventions over overlapping variable sets. J. Mach. Learn. Res., 16:2147-2205, 2015.","title":"Constraint-based causal discovery from multiple interventions over overlapping variable sets","authors":"S Triantafillou, I Tsamardinos","year":"2015","venue":"J. Mach. Learn. Res","doi":null,"url":null,"internal_url":null},{"position":39,"raw":"[Triantafillou et al., 2010] S. Triantafillou, I. Tsamardinos, and I. G. Tollis. Learning causal structure from overlap- ping variable sets. In AISTATS. JMLR, 2010.","title":"Learning causal structure from overlapping variable sets","authors":"S Triantafillou, I Tsamardinos, I G Tollis","year":"2010","venue":"AISTATS. JMLR","doi":null,"url":null,"internal_url":null},{"position":40,"raw":"[Yuan and Malone, 2013] C. Yuan and B. Malone. Learning optimal Bayesian networks: A shortest path perspective. J. Artif. Intell. Res., 48:23-65, 2013.","title":"Learning optimal Bayesian networks: A shortest path perspective","authors":"C Yuan, B Malone","year":"2013","venue":"J. Artif. Intell. Res","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}