Catálogo de publicaciones - libros

Maintenance of logical robustness in Information Integration represents a major challenge in the envisioned Semantic Web. In this framework, it is previsible unprecise information (with respect to an ontology) is retrieved from some resources. The sound integration of such information is crucial to achieve logical soundness. We present a approach to classify that knowledge by means of the of the possible robust ontology extensions and data.

In this paper, a first step toward the use of Artificial Intelligence tools (namely proof assistants) in the formal analysis of programs for Web services coordination is presented. This first attempt consists in the formal modeling of a system with transactional capabilities. The model is devised on a variant of the well-known Linda model for generative communication. We explore then the role of the Rete algorithm to implement efficiently a transactional operation, opening the way for a further formal analysis of it, by means of automated testing against a certified program (i.e. a program verified with the help of a proof assistant).

This paper is focused on description of a partial project of a multi-agent system for information retrieval. It is an integral part of a large ongoing project: the design of a multi-agent system for supporting medical diagnostics and monitoring in the area of cardiology (the ADAM system). Cooperating agents provide a very natural means of automating the pre-processing and (at least partially) the evaluation of a vast amount of medical data utilizing all available medical knowledge. As additional and very important functionality we have decided to design and implement agents that serve for building ontologies from medical resources on the Internet. The current phase of the project is targeted at implementing the system and putting it into full use.

Information integration plays a substantial role for data warehouses with their needs for highly dynamical adaptivity. For data and knowledge integration an OLAP–based method is discussed with special emphasis on summary level integration. The basic OLAP–model is enlarged by a so called composite structure for documentation of the concepts behind the data according to the needs of statistical information processing. The model supports also bookkeeping of the evolution of the data model.

Adaptation to the structure of the different University studies to the European Higher Education Area (EHEA) implies a deep reform affecting the structure of University studies. A brain storming about the mathematical courses will also be analysed.

After some considerations about e-learning and b-learning we will describe the main features of the project: : dMath. We pay attention to the general characteristics of the project and the different modules comprising it.

To foster the process of learning to program, theory and exercises are necessary. Traditionally, manual review of the source is used to provide feedback for the solutions. The effort is substantial and identified problems are prone to subjective interpretation. We study static analysis and testing tools as an enhancement to reviews and discuss the benefits. We analyze our findings by comparing the results from analysis by cross-checking the outcomes of the different tools with each other, with the tutors’ remarks, and with the course outcome additionally taking into account final examination results. The effort was substantial and it turned out, that the tools are no replacement for manual review. Tool support is an enhancement due to localization of problem areas, accurate check of programming rules, and an efficient way to detect plagiarism.

We present a model and a tool that allow the development of concept maps that can be navigated by the student, thereby allowing plenty of course contents to be browsed through a concept map. The tool applies several user adaptation techniques that reduce disorientation and cognitive overhead when browsing large maps by adapting the concept maps to the student’s knowledge. In addition, an evolutionary framework is used so consistency after changes is always guaranteed. The paper also describes a preliminary application to a real case at a University course.

Web-based learning environments are now extensively used. To guarantee the success in the learning processes, instructors require tools which help them to understand how these systems are used by their students, so that they can undertake more informed actions. Therefore, the aim of this paper is to show a Monitoring and Analysis Tool for E-learning Platforms (MATEP) which is being developed in the University of Cantabria (UC) to help instructors in these tasks. For this, web intelligence techniques are used.

Knowledge Representation & Reasoning (KR&R) is a fundamental topic in Artificial Intelligence. A basic KR language is First–Order Logic (FOL), the most representative logic–based representation language, which is part of almost any introductory AI course. In this work we present KRRT (Knowledge Representation & Reasoning Tutor). KRRT is a Web–based system which main goal is to help the student to learn FOL as a KR&R language.

The active semantic in Adaptive Object-Model (AOM) is only one example of the more complex active semantics at different orders. When we violate the integrity of the system, uncertainty grows up in the system. When monitors are conflicting agents in the modal logic, we can study uncertainty with the logic of the monitors that is comparable with the logic in the fuzzy set theory. Fuzzy values (integration degree) of a concatenation of interactive objects can be computed by fuzzy AND, OR and NOT operators presented in this paper. A short presentation of RDM framework suggests how it is possible to use object oriented language to implement adaptive object modelling. At the same time, relationship extension at different orders and also violation of the integrity of the system (uncertainty), where we break the constrain conditions; can be studied with RDM.