Efficient operations in feature terms using constraint programming

Abstract

Feature Terms are a generalization of first-order terms that have been introduced in theoretical computer science in order to formalize object-oriented capabilities of declarative languages, and which have been recently receiving increased attention for their usefulness in structured machine learning applications. The main obstacle with feature terms (as well as other formal representation languages like Horn clauses or Description Logics) is that the basic operations like subsumption have a very high computational cost. In this paper we model subsumption, antiunification and unification using constraint programming (CP), solving those operations in a more efficient way than using traditional methods. © 2012 Springer-Verlag Berlin Heidelberg.