Seminario “Combining Description Logics and Answer Set Programming for Modeling and Solving Concept Learning Problems” Dott.ssa Francesca Alessandra Lisi, 09.02.2017

Avviso di Seminario

 

“Combining Description Logics and Answer Set Programming for Modeling and Solving Concept Learning Problems”

 (Dott.ssa Francesca Alessandra Lisi)

 

Si terrà giorno 09.02.2017 alle ore 15:00 presso la Sala Seminari del DIMES (Cubo 42C – V Piano) il seminario dal titolo “Combining Description Logics and Answer Set Programming for Modeling and Solving Concept Learning Problems”, tenuto dalla Dott.ssa Francesca Alessandra Lisi del Dipartimento di Informatica dell’Università degli studi di Bari “Aldo Moro”.

 

Combining Description Logics and Answer Set Programming for Modeling and Solving Concept Learning Problems ==========================================================================================================

Research in Machine Learning (ML) has traditionally focussed on designing effective algorithms for solving particular tasks, most of which can be seen as Constraint Satisfaction Problems (CSPs). However, there is an increasing interest in providing the user with a means for specifying what the ML problem in hand actually is rather than letting him struggle to outline how the solution to that problem needs to be computed. This corresponds to a model+solver approach to ML, in which the user specifies the problem in a \emph{declarative modeling language} and the system automatically transforms such models into a format that can be used by a \emph{solver} to efficiently generate a solution. Concept Learning defines a class of ML problems where one such approach turns out to be viable. In this paper, we propose a model+solver approach to some Concept Learning problems which combines the efficacy of \emph{Description Logics} (DLs) in conceptual modeling with the efficiency of \emph{Answer Set Programming} (ASP) solvers in dealing with CSPs. In particular, the approach consists of a declarative modeling language based on second-order DLs under Henkin semantics, and mechanisms for enabling the evaluation of second-order DL formulas by ASP solvers.