Isambo Karali

Semantic search in the World News domain using automatically extracted metadata files

The Semantic Web can have great influence on various domains of information. One of them is the domain of World News. Semantic Web technologies aim at providing the means to organize the vast amount of knowledge that is scattered in the Web, in a machine understandable way. Then, searching and data retrieval would be much easier. This would be particularly helpful in the World News domain. There is a big variety of news sources and it would be useful to provide an efficient method to automatically organize them. In this paper, we describe World News Finder, a system which performs semantic search on the World News domain. The system is based on metadata files created for every single World News HTML webpage in an automatic way. According to a user query, the system performs the search on these metadata files rather than keyword search. To achieve the above, we developed the World News Ontology and a large set of domain-specific heuristic rules.

Agents for Querying Distributed Statistical Databases Over the Internet

Distributed database techniques and the Internet provide producers of statistics with a means to publish their data and metadata widely and make them available to a variety of users. Data matching to a user query and data access as well as data harmonization are some of the problems that should be solved. Intelligence is required in various stages of query answering and data matching. Moreover, the breadth and distributed nature of the Internet urge for a distributed approach. Agents seem to be the means by which both intelligence and distributed processing can be achieved. This paper presents a distributed approach for answering queries on statistical data that exist over the Internet using a multi-agent framework.

Logic Programming to Address Issues of the Semantic Web

The size of the Web and its increase rate made it cumber-some to locate high precision results to a requested piece of information. The Semantic Web provides a framework and a set of technologies enabling an effective machine processable information, aiming at a better computer-computer and human-computer communication. What if, though, we use a both machine processable and human understandable approach which can also apply in existent HTML Web sources? In this work, we investigate the problems being solved with the Semantic Web technologies and how this can be coped with Logic Programming techniques, especially Modular Logic Programming. We discuss issues from the data level, metadata and reasoning. Last but not least, we discuss agents. What is more is that, we claim that these techniques can be applied in the current Web information sources providing formal semantics for some aspects of the traditional Web.

MaTourA: multi-agent tourist advisor

MaTourA is a tourist advisory system about Greece that is being implemented in the parallel constraint logic programming language ElipSys. The purpose of MaTourA is to facilitate the work carried out in travel agencies by providing an interactive way to construct personalized tours, select predefined package tours and handle the underlying touristic information. The system has been designed as a set of high-level interacting agents. In this direction, the ElipSys language was extended with the appropriate features to support the development of multi-agent systems.

A versatile module system for Prolog mapped to flat prolog

This paper presents a module system for Frolog with large functionality, flexibility and various desirable properties. No restriction to the use of ordinary Prolog is imposed. On the contrary, the set of the relevant extra logical@icakts is extended so that to take the module concept into account. No space outside the modules exists and modules may be loaded and deleted at any time. The system is intended to be used either in cases where complicated interconnections are required or as a basis to build any module system structure on top of it, The complete functionality of the modute system is implemented as a preprocessor that maps modular programs to flat Prolog code. In addition, the implementation comprises the definition of a few run time support predicates. The considered mapping ean also seine as a specification of the module system and expsses the semanties of the introduced module support constructs, The implementation itself was tied out in Prolog.

PETINA—tour generation using the ElipSys inference system

PETINA is a Personalized Tourist Information Advisor system aiming at helping tourists to construct tours satisfying specified constraints. The system consults a large database that contains tourist data. PETINA has been implemented in the ElipSys language, which is a pure parallel logic programming system extended with various powerful mechanisms and features to allow efficient parallel execution. Although the expressive power of logic programming is profitable for the development of PETINA, standard Prolog systems lack of facilities that are vital for the PETINA application to work. ElipSys haa proved to be very suitable tool for the implementation of PETINA, as most of the former’s features are indispensable for handling the complexity of the encountered problems.

Uncertainty reasoning for the Semantic Web based on Dempster-Shafer model

The Semantic Web vision introduces the concept of machine-oriented web information. This information is characterized by the notions of uncertainty and vagueness. Towards this notions, an ontology model for representing uncertain or vague information, as well as an inference method for reasoning about this kind of information is necessary. Our approach faces the aforementioned issues, i.e. the imperfect ontology model and the reasoning approach. In order to show how our method performs, an implementation is outlined.

A Refinement of Import/Export Declarations in Modular Logic Programming and its Semantics

Encapsulation constructs with import/export declarations is the structuring facility offered in most commercial Prolog systems. However, real-life applications have shown to require a finer information exchange between encapsulated pieces of code. In this paper, a refinement of import/export declarations for modules of logic programs is presented. This offers a stricter form of communication between the modules and a larger variety of visibility states of their predicates, the standard approaches being special cases of it. The semantics of this module system has been examined and model-theoretic, fixpoint and operational ones are given and have been proved to be equivalent. Instead of using other logics, all these semantics extend the ones of Horn clause logic using concepts commonly used in it. In addition, the module system has been naturally transformed to Horn clause logic exploiting the distinction of the predicates within a module according to the interface declarations of this module. A form of equivalence with the other semantics of the system is given. In addition, the employed transformation has provided us with a basis for a preprocessor based implementation of the module system.

A fuzzy logic approach for reasoning under uncertainty and vagueness - a matchmaking case study

Matchmaking problems have been studied extensively in literature. Todays World Wide Web provides us with the ability to have access to a vast amount of data. The manual matchmaking cannot be considered a feasible solution and automatic methods by machines have been evolved. The matchmaking task is always based on a set of criteria, but these criteria may not have a well-defined meaning. In this paper, a matchmaking method of web data based on fuzzy criteria is outlined.

Uncertainty reasoning for the "big data" semantic web

The Semantic Web introduces the concept of machine-oriented information, i.e. information that can be processed by machines or agents without human intervention. In order to achieve this, web information should be represented in a way that its semantics is understandable by agents. Defining semantics for web information is not an easy process, as the web information is not always clear-cut. For example, a web search for comfortable hotels introduces the vague concept comfortable. So, semantics are always related to some kind of vagueness. Moreover, the source of web information is always characterized by a notion of uncertainty, e.g Ninety percent of four star hotels have a swimming pool. Uncertainty and vagueness can be strongly related and this relation demands an extension of any representation scheme in order to capture imperfect concepts. Towards this notion we propose an ontology as well as a reasoning method suitable for imperfect data.