Abstract
This paper studies the problem of learning parameters for global constraints such as SEQUENCE from a small set of positive examples. The proposed technique computes the probability of observing a given constraint in a random solution. This probability is used to select the more likely constraint in a list of candidates. The learning method can be applied to both soft and hard constraints.
Author supplied keywords
Cite
CITATION STYLE
Picard-Cantin, É., Bouchard, M., Quimper, C. G., & Sweeney, J. (2016). Learning parameters for the sequence constraint from solutions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9892 LNCS, pp. 405–420). Springer Verlag. https://doi.org/10.1007/978-3-319-44953-1_26
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
