Catálogo de publicaciones - libros

In this paper we present a logic programming based framework for the integration of databases. Our framework is based on a logic language obtained by extending Datalog with several constructs such as choice, set constructors, aggregates. We show that the extended language can be profitable used to easily implement most of the integration techniques defined in the literature and define more general techniques.

The inference rule ℧-resolution was introduced in [] as a technique for developing an SLD-style query answering procedure for the logic programming subset of annotated logic. This paper explores the properties of ℧-resolution in the general theorem proving setting. In that setting, it is shown to be complete and to admit a linear restriction. Thus ℧-resolution is amenable to depth-first control strategies that require little memory. An ordering restriction is also described and shown to be complete, providing a promising saturation-based procedure for annotated logic. The inference rule essentially requires that the lattice of truth values be . Earlier investigations left open the question of whether all distributive lattices are ordinary; this is answered in the affirmative here.

We argue for a semantical modification of the language DLP. We show by examples that the current DLP representation in some cases does not provide intuitive answers, in particular when applied to inheritance reasoning. We present and discuss an initial modification of DLP that yields the expected answers in some examples that we consider significant

Some new types of knowledge constructions in epistemic logic are defined, and semantics given by combining game-theoretic notions with modal models. One such notion introduced is focussed knowledge, which arises from imperfect information in quantified epistemic logics. This notion is useful in knowledge representation schemes in artificial intelligence and multi-agent systems with uncertainty. In general, in all the logics considered here, the imperfect information is seen to give rise to partiality, including partial common and partial distributed knowledge. A game-theoretic method of creating non-monotonicity will then be suggested, based on the partialised notion of ‘only knowing’ and inaccessible possible worlds. The overall purpose is to show the extent in which games combine with a given variety of knowledge constructions.

Two properties characterize the quality of a knowledge base (KB): coherency and completeness. In our work, the completeness of the KB is evaluated on a set of test cases. The incompleteness is revealed by test cases that cannot be proved, called deficiencies. We also use test cases to check the coherency of the KB: we propose a new notion, called C_coherency, that extends previous definitions of coherency. The situations that reveal the C_incoherency of the KB are called conflicts. When the problems in a KB have been identified, our aim is to restore the validity of this base. For each conflict and each deficiency, we determine which modifications of the KB can suppress this problem without introducing new deficiencies or conflicts in the KB.

One of the most important problems on rule induction methods is that they cannot extract rules, which plausibly represent experts’ decision processes. On one hand, rule induction methods induce probabilistic rules, the description length of which is too short, compared with the experts’ rules. On the other hand, construction of Bayesian networks generates too lengthy rules. In this paper, the characteristics of experts’ rules are closely examined and a new approach to extract plausible rules is introduced, which consists of the following three procedures. First, the characterization of decision attributes (given classes) is extracted from databases and the classes are classified into several groups with respect to the characterization. Then, two kinds of sub-rules, characterization rules for each group and discrimination rules for each class in the group are induced. Finally, those two parts are integrated into one rule for each decision attribute. The proposed method was evaluated on a medical database, the experimental results of which show that induced rules correctly represent experts’ decision processes.

Our research has been focused on developing techniques for solving binary constraint satisfaction problems (CSP) using evolutionary algorithms, which take into account the constraint graphs topology. In this paper, we introduce a new idea to improve the performance of evolutionary algorithms, that solve complex problems. It is inspired from a real world observation: The ability to evolve for an individual in an environment that changes is not only related to its genetic material. It also comes from what has learned from it parents. The key idea of this paper is to use its inheritance to dynamically improve the way the algorithm creates a new population using a given set of operators. This new dynamic operator selection strategy has been applied to an evolutionary algorithm to solve CSPs, but can be easily extended to other class of evolutionary algorithms. A set of benchmarks shows how the new strategy can help to solve large NP-hard problems with the 3-graph coloring example.

As an increasing number of Web sites consist of an increasing number of pages, it is more difficult for the users to rapidly reach their own target pages. So the intelligent systems supporting the users in navigation of the Web contents are in high demand. In this paper, we describe an intelligent recommendation system called the system L-R, which constructs user models by mining the Web access logs and recommends the relevant pages to the users based both on the user models and the Web contents. We have evaluated the prototype system and have obtained the positive effects.

We describe the design and algorithms , a scholarly publication retrieval system. This system is intended to meet library user studies which advocate human interpretation and social interactions. The metaphors we used are annotations and publication (making public). We first discuss about different philosophical approaches to semantics and choose the more suited to scholarly work: the one considering a transitory, hypothetical and polemical knowledge construction. Then we propose an overview of : an hypertext system based on a dynamic structure of user annotations. The visualization and evolution of the structure (a dynamic directed acyclic graph) is made more efficient by the use of an browsing algorithm.

In manufacturing fabricated sheet metal parts, the required shape has to be bent from the flat 2-D layouts. In this bending process, the most complex and critical work is determining the bend sequence and assigning appropriate tools for each bend. Determining the bend sequence is itself a combinatorial problem and this when coupled with tool assignment leads to a huge combination and clearly shows an exhaustive approach is impossible and we propose Genetic Algorithm (GA), an adaptive algorithm to solve the problem. Information regarding the operator knowledge and operator desire are input to the system to generate efficient bending process. And moreover, in order to improve the performance of GA, a preprocessor is being implemented which searches combinable bends and thereby reduce search space and solve the problem in time-economic way.