This chapter suggests that logic programs employing recent advances in semantics and in knowledge representation formalisms provide a more robust framework in which to develop very intelligent systems in any domain of knowledge or application. The author has performed work applying this paradigm and these reasoning formalisms in the areas of financial applications, security applications, and enterprise information systems.
TopBackground
This is a primer for computing professionals, and not an authoritative work on the state-of-the-art of Logic Programming. Logic programming presents us with an excellent tool to develop a variety of powerful intelligent systems. What will be discussed is a particular version of logic programming called “ASP”, which means “Answer Set Programming”. “Version” in this context is defined by the semantics of the language, and not by performance issues. ASP is currently the most popular and the most well researched Logic Programming language. (To demonstrate the importance that ASP is receiving in the field, an entire issue of “AI Magazine” has been devoted to ASP. This journal is the flagship journal for the Association for the Advancement of Artificial Intelligence, the premier scholarly and practitioner association in the field of Artificial Intelligence. The issue mentioned is the most current issue at the time of this writing, Volume 37 Number 3, Fall 2016.)
While there are still issues that need to be addressed in Logic Programming, and while there may be additional non-logical techniques to complement logic-based systems, it is reasonable to believe that logic will form the cornerstone of any serious machine intelligence in the future. Consider that the goal of the field of artificial intelligence (in fact, one of the early goals of computing itself) is to build “HAL”, the all-knowing, self-sufficient computer of science-fiction movies (Clarke, 1968). To this end, it behooves us to understand, use, and further refine this paradigm.
In this paper, the syntax and of semantics of ASP shall be presented. The early forms of this language were initially called the Stable Model Semantics, and later, Disjunctive Logic Programming (Gelfond & Lifschitz, 1988; Gelfond & Lifschitz 1991). This language is a purely declarative language (as opposed to procedureal languages which dominate industry, and as opposed to functional languages which are the main competitor in the field of AI.) Declarative languages have their roots in logic programming (Kowalski, 1974; Kowalski, 1979), they are defined in the syntax and semantics of standard Prolog (Colmerauer, et. al., 1973; Clark 1978), and they employ the work on nonmonotonic logic (Reiter, 1980; Moore, 1985).