Semantic technologies based on logic programming

This thesis describes work on how Logic Programming (LP) can be used in various knowledge intensive applications. The contributions described in the thesis include algorithms and tools for efficient Description Logics reasoning, semantic comparison of source programs and semi-automated creation of metainformation. The link between these results is that (1) they all introduce novel semantic technologies, (2) they apply logic programming techniques in general and the Prolog LP language in particular, and (3) they have been actually implemented in the form of various applications. Semantic technologies form the basis of building software systems that work with the meaning of various kinds of information rather than with their textual form. The main idea is to reason about some sort of knowledge represented in a machine processable way. The most important parts of this process are knowledge representation and knowledge management. The dissertation deals with two groups of knowledge representation and management technologies: Logic Programming and Description Logic systems. Combination of these approaches creates interesting challenges: the idea of the so-called Description Logic Programming has been introduced in 2003. Several results in the present dissertation contribute to this area of semantic technologies. In the context of Description Logics we have addressed two issues. The first issue is about how to efficiently reason on Description Logic knowledge bases when huge amounts of data are present. We have created the theoretical basis of a new Description Logic reasoning system, called DLog, that transforms description logic axioms into a Prolog program. This transformation is done independently from the individuals: they are accessed dynamically during the Prolog execution of the generated program. For this we have created several algorithms and proved their useful properties. As a result, with our implementation, we reached better scalability and more efficient execution compared to that of the earlier approaches. The second issue addresses the question of how to use conceptual models for information integration. We have created a transformation framework together with a modelling methodology that uses Prolog for querying conceptual models formulated in Description Logics. Here we apply the closed world assumption as we argue it fits the context of information integration better. We demonstrate this approach by applying it for the SINTAGMA information integration system. The thesis also discusses two use cases that present the viability of Logic Programming for knowledge intensive applications. First, we deal with the issue of how to detect plagiarism between source programs. We have designed and implemented Match, a generic plagiarism detection framework that aims at the semantic comparison of source programs. The idea is to transform the source code into mathematical objects, use appropriate reduction and comparison methods on these, and interpret the results appropriately. We have been using the Prolog implementation of the system at the Budapest University of Technology and Economics to successfully detect plagiarism in homework assignments for the past eight years. Finally, we investigate the question of how to deduce meta-information in systems working with various kinds of documents. We have developed the SREngine framework, which manages a pool of generic objects together with their properties and supports Prolog based reasoning on these. The idea is to infer new properties about the objects, using their existing properties and a set of user-defined rules. These rules are given in a special logic based language developed for SREngine.