Epistemic Answer Set Programming

Ezgi Iraz Su (1, 2)

1 Istinye University
2 IRIT-LILaC - Logique, Interaction, Langue et Calcul

This paper introduces a new epistemic extension of answer set programming (ASP) called epistemic ASP (E-ASP). Then, it compares E-ASP with existing approaches, showing the advantages and the novelties of the new semantics and discusses which formalisms provide more intuitive results: compared to Gelfond’s epistemic specifications (ES), E-ASP defines a simpler, but sufficiently strong language. Its epistemic view semantics is a natural and more standard generalisation of ASP’s original answer set semantics, so it allows for ASP’s previous language extensions. Moreover, compared to all semantics proposals in the literature, epistemic view semantics facilitates understanding of the intuitive meaning of epistemic logic programs and solves unintended results discussed in the literature, especially for epistemic logic programs including constraints.

Mots clés: Answer set programming; Epistemic specifications; Modal logic S5; Stable models; Answer sets; World views; Autoepistemic equilibrium models

Domaines: Intelligence artificielle [cs.AI]; Logique en informatique [cs.LO]

Ezgi Iraz Su. Epistemic Answer Set Programming. European Conference on Logics in Artificial Intelligence (JELIA 2019), May 2019, Rende, Italy. pp.608-626. ⟨hal-02421559⟩