Journal:Informatica
Volume 9, Issue 2 (1998), pp. 161–171
Abstract
In this paper we show that the least mean square (LMS) algorithm can be speeded up without changing any of its adaptive characteristics. The parallel LMS adaptive filtering algorithm and its modifications are presented. High speed is achieved by increasing the parallelism in the LMS adaptive algorithm through a proper modification of the LMS adaptive algorithm. An iterative procedures for efficient computation of the lower triangular inverse matrix and the input signal covariance matrix are presented.
Journal:Informatica
Volume 9, Issue 2 (1998), pp. 141–160
Abstract
The nonlinearities play a crucial role in the brain processes. They take place in neuronal system elements: synapses, dendrite membranes, soma of neurons, axons. It is established that the soma nonlinearity, which is of sigmoidal shape, is not so strong as compared with the electric current-voltage relation of a dendrite membrane. The relation is N-shaped with two stable and one unstable points. In dynamics, this leads to the appearance of a switch wave or formation of some logic functions. We present some artificial logic circuits based on an electrical analogy of dendritic membrane characteristics in static and dynamic cases. The nonlinear cable theory and the numerical simulation were used. Basing on the logic circuit construction proposed, we suppose that the dendritic membrane processes are able not only to gather and transfer information but also to transform and classify knowledge.
The theoretical substantiation and numerical experiments are only the first step forward to the proving of neuronal dendritic logic constructions. Of course, extensive neurophysiological tests are necessary to discover the final mechanism of neuronal computing in the human brain.
Journal:Informatica
Volume 9, Issue 2 (1998), pp. 123–140
Abstract
In this paper we describe implementation of numerical adaptive algorithms for multi-dimensional quadrature on distributed-memory parallel systems. The algorithms are targeted at clusters of workstations with standard message passing interfaces, e.g., PVM or MPI. The most important issues are communication and load balancing. Static and dynamic partitioning of the region are considered. Numerical results on various workstation clusters are reported.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 107–117
Abstract
This paper presents the declarative extension of the deductive database system LOLA to the object-oriented deductive database system O!-LOLA. The model used for O!-LOLA is “objects as theories”, extended by state evolution. O!-LOLA combines logic programming and OO programming in two different ways: First, methods are implemented as logic programs. These methods can be inherited, encapsulated and overloaded. Second, logic programs can be defined over classes, meta-classes, instances, attributes and values. Dynamic updates of attributes of objects and dynamic instantiations of classes are supported.
O!-LOLA is implemented as a preprocessor. O!-LOLA programs are transformed into LOLA rules and facts, which are evaluated set-oriented and bottom-up, using fixpoint semantics. Some object-oriented features concerning dynamic aspects are handled via built-in predicates in LOLA.
We describe the applied theory, the system and the preprocessor, including an example of how methods are translated and we discuss dynamic updates of objects in O!-LOLA.
The benefits of our system in contrast to others are: a single integrated language, clear semantics and a set-oriented evaluation. O!-LOLA uses fixpoint semantics (not any procedural semantics like other systems) and still evaluates set-oriented (and not in a mixed manner like other systems). Thus, we can fully use all optimization techniques developed for deductive databases and gain a very efficient system.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 85–105
Abstract
Attributed automaton (AA) is a formalism for conceptual knowledge specification using regular syntax with attributes representing contextual relations as well as semantic properties of concepts. AA can be treated as a generalization of a finite automaton with attributes and computational relations attached to states and transitions respectively. In this paper we develop a new specification method for AA based on functional combinators. It allows modular specification of AA, enjoys good algebraic properties and is extendable for different kind of attributed automata.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 65–84
Abstract
Statelog is a Datalog extension integrating the declarative semantics of deductive rules with the possibility to define updates in the style of active and production rules. The language is surprisingly simple, yet captures many essential features of active rules. After reviewing the basics of active rules, production rules, and deductive rules, we elaborate on the problem of handling rule termination in the context of Statelog: It is undecidable whether a Statelog program terminates for all databases, and PSPACE-complete for a given database. The latter can be accomplished within the logical language: for every Statelog program P, there is a terminating program P↓ which decides for any given database 𝒟, whether P ∪ 𝒟 terminates.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 51–64
Abstract
An extention of the QBE language in the form of graph queries is proposed for databases having complex schemata. These graph queries provide user interface for the ER model, like QBE queries provide user interface for the relational model. The implementation and usage of graph queries of ER-based “MicroPoisk” DBMS are presented.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 21–50
Abstract
In this paper we examine data dependence in a nested loop programs which are obtained by inserting one loop program into another. This is viewed as the composition of structural modules (S-modules) in the structural blanks (SB) approach. SB is a method for expressing computations based on recurrence relations. It is built on top of traditional programming languages like Fortran or Pascal. SB aims at supporting the transformational development and reuse of program modules that have complex data dependence patterns and provides an architectural framework for software packages.
Journal:Informatica
Volume 9, Issue 1 (1998): Special Issue on Programming Theory, Information System Engineering, Software Engineering, and Artificial Intelligence, pp. 5–20
Abstract
In the paper, we describe our approach and experience with teaching fundamentals of functional programming by program schemata construction and explanation. Program schemata for processing of lists are presented. Our approach reflects our ultimate goal – to support the learning process. As the main result, we report on experiments that allow to judge quite favourably the approach to teaching functional (Lisp) programming when the student learns a set of program schemata and how to apply them.
Journal:Informatica
Volume 8, Issue 4 (1997), pp. 599–605
Abstract
Let G0 and G1 be arbitrary fuzzy classifiers (Vatlin, 1993). We say that G1 improves G0 if the performance of G1 is more than G0 one. We also introduced the concepts of consistent and strongly selfguessing fuzzy classifiers. The criterion of strong selfguessing is formulated. The theorems on the conditions of probabilistic improvement of consistent and monotonic improvement of strongly selfguessing fuzzy classifiers are proved.