Extrapolating constraint networks by symbolic classification

Simple item page
dc.contributor.authorPrestwich, Steven D.
dc.contributor.funderScience Foundation Irelanden
dc.contributor.funderEuropean Regional Development Funden
dc.date.accessioned2022-06-08T14:15:00Z
dc.date.available2022-06-08T14:15:00Z
dc.date.issued2022-07
dc.date.updated2022-06-08T14:05:42Z
dc.description.abstractConstraint acquisition (CA) methods aim to learn constraint satisfaction problems (CSPs) from data, thus automating the difficult and error-prone task of constraint modelling. Most CA methods learn CSPs of a certain size from examples of that size, for example an 8-queens CSP from 8-queens solutions. A few methods can learn a CSP from solutions of other sizes, or learn a generalised CSP for all sizes, which is more challenging but also more useful in practice. This paper describes a new approach to learning a CSP: extrapolation from learned CSPs of other sizes. This has the advantage that the training CSPs can be learned by any convenient CA method, thus potentially handling positive-only, negative-only, positive-negative, noisy or unlabelled data. We model the problem as classification, and show that a symbolic classifier based on genetic programming outperforms alternatives such as random forests and deep learning.en
dc.description.sponsorshipScience Foundation Ireland (Grant No. 12/RC/2289-P2 co-funded under the European Regional Development Fund); Science Foundation Ireland (CONFIRM Centre for Smart Manufacturing, Research Code 16/RC/3918)en
dc.description.statusPeer revieweden
dc.description.versionAccepted Versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.citationPrestwich, S. D. (2022) 'Extrapolating Constraint Networks by Symbolic Classification ', IJCAI 2022 DSO Workshop: Data science meets optimisation, Messe Wien, Vienna, Austria, 23-29 July, forthcoming publication.en
dc.identifier.endpage6en
dc.identifier.startpage1en
dc.identifier.urihttps://hdl.handle.net/10468/13288
dc.language.isoenen
dc.publisherIJCAIen
dc.relation.ispartofIJCAI 2022 DSO Workshop
dc.relation.projectinfo:eu-repo/grantAgreement/SFI/SFI Research Centres/12/RC/2289/IE/INSIGHT - Irelands Big Data and Analytics Research Centre/en
dc.rights© 2022 IJCAIen
dc.subjectConstraint acquisition (CA)en
dc.subjectConstraint satisfaction problems (CSPs)en
dc.subjectConstraint optimisationen
dc.subjectData scienceen
dc.subjectOperations researchen
dc.subjectOptimisationen
dc.titleExtrapolating constraint networks by symbolic classificationen
dc.typeConference itemen
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
dso.pdf
Size:
165.3 KB
Format:
Adobe Portable Document Format
Description:
Accepted version
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
2.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Computer Science - Conference Items
Insight Centre for Data Analytics - Conference Items