{"type":"ScholarlyArticle","title":"Improved Acyclicity Reasoning for Bayesian Network Structure Learning with Constraint Programming","authors":[{"name":"Fulya Trösser","slug":"fulya-trosser","given":"Fulya","family":"Trösser","affiliation":""},{"name":"Simon de Givry","slug":"simon-de-givry","given":"Simon","family":"de Givry","affiliation":""},{"name":"George Katsirelos","slug":"george-katsirelos","given":"George","family":"Katsirelos","affiliation":""}],"abstract":"Bayesian networks are probabilistic graphical models with a wide\r\nrange of application areas including gene regulatory networks inference, risk\r\nanalysis and image processing. Learning the structure of a Bayesian\r\nnetwork (BNSL) from discrete data is known to be an NP-hard task with\r\na superexponential search space of directed acyclic graphs.  In this\r\nwork, we propose a new polynomial time algorithm for discovering a\r\nsubset of all possible cluster cuts, a greedy algorithm for\r\napproximately solving the resulting linear program, and a generalized\r\narc consistency algorithm for the acyclicity constraint.  We embed\r\nthese in the constraint programming-based branch-and-bound solver\r\nCPBayes and show that, despite being suboptimal, they improve\r\nperformance by orders of magnitude. The resulting solver also compares\r\nfavorably with GOBNILP, a state-of-the-art solver for the BNSL\r\nproblem which solves an NP-hard problem to discover each cut and\r\nsolves the linear program exactly.","doi":"10.24963/ijcai.2021/584","url":"https://ijcai.org/proceedings/2021/584","pdf":"https://ijcai.org/proceedings/2021/0584.pdf","markdown":"https://ijcai.org/proceedings/2021/584.md","bibtex":"https://ijcai.org/proceedings/2021/bibtex/584","volume":{"slug":"2021","title":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2021","url":"https://ijcai.org/proceedings/2021/","year":2021,"location":"Montreal","dates":"19-27 August 2021","editors":"Prof. Zhi-Hua Zhou","isbn":"978-0-9992411-9-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2021"},"section":"Uncertainty in AI","track":"Main Track","number":584,"pages":{"first":4250,"last":4257},"keywords":["Uncertainty in AI: Bayesian Networks","Constraints and SAT: Constraint Optimization"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}