Search: Show all articles

Multiconference Proceedings (PDF, 55.773 M)

Search results for

Workshop on Agent Based Computing V

• Agent architecture for building Robocode players with SWI-Prolog
  142 agent architecture, agent environment, robocode, prolog Vasile Alaiba, Armand Rotaru, pages 3 – 7. Show abstract Effective study of autonomous agents is a challenging activity, for researchers and programmers alike. In this paper we introduce a new agent environment built on top of the tank-simulation game Robocode. We consider this game a good choice for testing agent programming strategies and an excellent learning opportunity. We integrated SWI-Prolog into Robocode and built an architecture for writing agents with Prolog. Game generated events are translated by our framework into predicate invocations, allowing the agent to perceive the environment. An API to issue commands to the tank from Prolog acts as agent's effectors. We tested our architecture by implementing a simple autonomous agent that uses a static strategy.
• Tackling Complexity of Distributed Systems: towards an integration of Service-Oriented Computing and Agent-Oriented Programming
  91 Agent oriented programming, Service Oriented Computing Giacomo Cabri, Letizia Leonardi, Raffaele Quitadamo, pages 9 – 15. Show abstract The development of distributed systems poses different issues that developers must carefully take into consideration. Web Services and (Mobile) Agents are two promising paradigms that are increasingly exploited in distributed systems design: they both try, albeit with very different conceptual abstractions, to govern unpredictability and complexity in wide-open distributed scenarios. In this paper, we compare the two approaches with regard to different aspects. Our aim is to provide developers with critical knowledge about the advantages of the two paradigms, stressing also the need for an intelligent integration of the two approaches.
• Connecting Methodologies and Infrastructures in the Development of Agent Systems
  37 Methodologies, infrastructures, meta-model Giacomo Cabri, Mariachiara Puviani, Raffaele Quitadamo, pages 17 – 23. Show abstract In the building of agent systems developers can be supported by both appropriate methodologies and infrastructures, which guide them in the different phases of the development and provide useful abstractions. Nevertheless, we assist to a situation in which methodologies and infrastructures are not connected each other: the products of the analysis and design phases could not always be exploited in the implementation phase in a direct way, even if sometimes CASE-tools are present to help in translating methodologies' diagrams in infrastructures' code. This leads to a “gap” between methodologies and infrastructures that is likely to produce fragmented solutions and to make the application maintenance difficult. In this paper we face this issue, proposing three directions to solve the problem. We do not want to propose a “new brand” methodology and infrastructure tightly connected, rather, we aim at reusing as much as possible what already exists, not only in abstract terms, but also in concrete “fragments” of methodologies; an appropriate meta-language that describes how a methodologies works would be useful to more easily map them onto the infrastructures, or even to “compose” a new methodologies. A further approach is based on an “intermediate” layer between methodologies and infrastructures, which provides a mapping between the involved entities.
• Adaptation of Extended Polymorphic Self-Slimming Agent Model into e-Sourcing Platform
  146 e-supply chains, e-commerce, e-marketplace, e-sourcing, agent technology Konrad Fuks, Arkadiusz Kawa, Waldemar Wieczerzycki, pages 25 – 29. Show abstract There are two main contributions of the work pre­sented in this paper. First, extended PSA agent model is descri­bed. It is based on two new properties of so called bootstrap agent: multiplication and parallelism. Second contribution is application of extended PSA model to e-sourcing platform. This duo (e-sourcing and extended PSA model) shows enterprises can significantly increase resources acquisition and potential suppliers search.
• The Triple Model: Combining Cutting-Edge Web Technologies with a Cognitive Model in a ECA
  208 agent architecture, cognitive modeling Maurice Grinberg, Stefan Kostadinov, pages 31 – 37. Show abstract This paper introduces a new model which is intended to combine the power of a connectionist engine based on fast matrix calculation, RDF based memory and inference mechanisms, and affective computing in a hybrid cognitive model. The model is called Triple and has the following three parts: a reasoning module making advantage of RDF based long-term memory by performing fast inferences, a fast connectionist mapping engine, which can establish relevance and similarities, including analogies, and an emotional module which modulates the functioning of the model. The reasoning engine synchronizes the connectionist and the emotional modules which run in parallel, and controls the communication with the user, retrieval from memory, transfer of knowledge, and action execution. The most important cognitive aspects of the model are context sensitivity, specific experiential episodic knowledge and learning. At the same time, the model provides mechanisms of selective attention and action based on anticipation by analogy. The inference and the connectionist modules can be optimized for high performance and thus ensure the real-time usage of the model in agent platforms supporting embodied conversational agents.
• Planning and Re-planning in Multi-actors Scenarios by means of Social Commitments
  188 social commitments, distributed planning, HTN, BDI Antonín Komenda, Michal Pěchouček, Jiří Bíba, Jiří Vokřínek, pages 39 – 45. Show abstract We present an approach to plan representation in multi-actors scenarios that is suitable for flexible replanning and plan revision purposes. The key idea of the presented approach is in integration of (i) the results of an arbitrary HTN (hierarchical task network) -oriented planner with (ii) the concept of commitments, as a theoretically studied formalism representing mutual relations among intentions of collaborating agents. The paper presents formal model of recursive form of commitments and discusses how it can be deployed to a selected hierarchical planning scenarioThe presented research has been supported by I-GLOBE (http://i-globe.info/), the US ARMY, CERDEC project (no.: N62558-06-P-0353). The project has been performed in cooperationwith University of Edinburgh, Artificial Intelligence Application Institute .
• Utility-based Model for Classifying Adversarial Behaviour in Multi-Agent Systems
  186 multi-agent systems, agent modelling, agent interactions, adversarial modelling Viliam Lisý, Michal Jakob, Jan Tožička, Michal Pěchouček, pages 47 – 53. Show abstract Interactions and social relationships among agents are an important aspect of multi-agent systems. In this paper, we explore how such relationships and their relation to agent's objectives influence agent's decision-making. Building on the framework of stochastic games, we propose a classification scheme, based on a formally defined concept of interaction stance , for categorizing agent's behaviour as self-interested, altruistic, competitive, cooperative, or adversarial with respect to other agents in the system. We show how the scheme can be employed in defining behavioural norms, capturing social aspects of agent's behaviour and/or in representing social configurations of multi-agent systems.
• A Case Study in Agent-Based Systems-Implementation: Distributed Information Management at the Public Prosecution
  119 distributed information management agents security Martijn Warnier, Reinier Timmer, Michel Oey, Frances Brazier, Anja Oskamp, pages 55 – 61. Show abstract Securely managing shared information in distributed environments across multiple organizations is a challenge. Distributed information management systems designed to this purpose, must be able to support individual organizations' information policies whilst ensuring global consistency and completeness of information. This paper describes a multi-agent based prototype implementation of a distributed information management system, for distributed digital criminal dossiers. The prototype implementation runs on the multi-agent platform AgentScape.

3rd International Symposium Advances in Artificial Intelligence and Applications

• Applying Differential Evolution to the Reporting Cells Problem
  117 Differential Evolution, Location Management, Reporting Cells, Mobile Networks Sónia Almeida-Luz, Miguel Angel Vega-Rodríguez, Juan António Gómez-Pulido, Juan Manuel Sánchez-Pérez, pages 65 – 71. Show abstract The Location Management problem corresponds to the management of the network configuration with the objective of minimizing the costs involved. It can be defined using several different schemes that principally consider the location update and the paging costs. The Location Area and Reporting Cells are two common strategies used to solve the location management problem. In this paper we present a new approach that uses the Differential Evolution algorithm applied to the reporting cells planning problem, with the objective of minimizing the involved location management costs. With this work we want to define the best values to the differential evolution parameters and respective scheme, using 12 distinct test networks, as well as compare our results with the ones obtained by other authors. The final results obtained with this approach are very encouraging.
• Towards Word Sense Disambiguation of Polish
  162 word sense diambiguation, machine learning Dominik Baś, Bartosz Broda, Maciej Piasecki, pages 73 – 78. Show abstract We compare three different methods of Word Sense Disambiguation applied to the disambiguation of a selected set of 13 Polish words. The selected words express different problems for sense disambiguation. As it is hard to find works for Polish in this area, our goal was to analyse applicability and limitations of known methods in relation to Polish and Polish language resources and tools. The obtained results are very positive, as using limited resources, we achieved the accuracy of sense disambiguation greatly exceeding the baseline of the most frequent sense. For the needs of experiments a small corpus of representative examples was manually collected and annotated with senses drawn from plWordNet. Different representations of context of word occurrences were also experimentally tested. Examples of limitations and advantages of the applied methods are discussed.
• USERING. Educational self-tuning–recommendations in the 8th level of ISO/OSI model.
  147 IT security, computational intelligence, usering Miroslaw Bedzak, pages 79 – 82. Show abstract It is autumn 2012…The VMware Infrastructure …3, 4 editions virtualised crucial components of IT environ­ment, i.e. computing (CPU, RAM), networking and storage. However, an important element was overlooked. Which one? A user. There was no mechanism built in into VI3/VI4 that would support administrator in gaining effectively the skills of imple­menting solutions advised by manufacturer, the so-called rec­ommendations (“best practice”, etc.). Either, the high level of user’s skills (primarily, administrator’s ones) were not treated as a valuable resource of LAN infrastructure that could be (should be) used, while a “surplus” could be virtualised for a common good of the society (local and/or global), concentrated around the enterprise-class infrastructure virtualisation tech­nology. The latest edition, VI5 beta, brings also an important change in this respect in the form of a new module, VMware Usering, which is directed in the current version (i.e. beta ver­sion) first of all towards security hardening, i.e. the linguistic inference has been accomplished by a method of fuzzy control basing on the knowledge-base built on the recommendation of VI5 beta manufacturer. The VMware Usering is a set of tools, stimulating users to take up actions being consistent with manu­facturer’s recommendations (VMware User Hardening) and virtualising (optionally for administrators with HIGH FUZZY rat­ing) the competence resource of advanced users (VMware User Competence Sharing). The rise of a resistance level of IT infra­structure to increasing threats of internet security is also a manu­facturer’s own interest, therefore one can expect in the near future a popularisation of VMware Usering-class solutions as an important tool supporting an average internet surfer in his/her solitary struggle against crackers at the level of 8th OSI model layer .
• Sense-Based Clustering of Polish Nouns in the Extraction of Semantic Relatedness
  161 Bartosz Broda, Maciej Piasecki, Stanisław Szpakowicz, pages 83 – 89. Show abstract The construction of a wordnet from scratch requires intelligent software support. An accurate measure of semantic relatedness can be used to extract groups of semantically close words from a corpus. Such groups help a lexicographer make decisions about synset membership and synset placement in the network. We have adapted to Polish the well-known algorithm of Clustering by Committee, and tested it on the largest Polish corpus available. The evaluation by way of a plWordNet -based synonymy test used Polish WordNet, a resource still under development. The results are consistent with a few benchmarks, but not encouraging enough yet to make a wordnet writer's support tool immediately useful.
• An immune approach to classifying the high-dimensional datasets
  141 negative selection, anomaly detection Andrzej Chmielewski, Sławomir Wierzchoń, pages 91 – 96. Show abstract This paper presents an immune-based approach to problem of binary classification and novelty detection in high-dimensional datasets. It is inspired by the negative selection mechanism, which discriminates between self and nonself elements using only partial information. Our approach incorporates two types of detectors: binary and real-valued. Relatively short binary receptors are used for primary detection, while the real valued detectors are used to resolve eventual doubts. Such a hybrid solution is much more economical in comparison with “pure” approaches. The binary detectors are more faster than real-valued ones, what allows minimize computationally and timely complex operations on real values. Additionally, regardless of type of encoding, the process of sample's censoring is conducted with relatively small part of its attributes.
• Application of Clustering and Association Methods in Data Cleaning
  111 data cleaning, data quality, data mining, attribute correction Lukasz Ciszak, pages 97 – 103. Show abstract Data cleaning is a process of maintaining data quality in information systems. Current data cleaning solutions require reference data to identify incorrect or duplicate entries. This article proposes usage of data mining in the area of data cleaning as effective in discovering reference data and validation rules from the data itself. Two algorithms designed by the author for data attribute correction have been presented. Both algorithms utilize data mining methods. Experimental results show that both algorithms can effectively clean text attributes without external reference data.
• A rule-based approach to robust granular planning
  153 granular planning, route planning, rule-based systems Sebastian Ernst, Antoni Ligęza, pages 105 – 111. Show abstract Real-time execution of planned routes often requires re-planning, especially in dynamic, highly unpredictable environments. Ad-hoc re-planning in case of failure of an optimal (suboptimal) plan leads to deterioration of solution quality. Moreover, it becomes costly and often unmanageable under real time constraints. This paper describes an approach of a partitioning scheme for urban route planning problem domain into a hierarchy of subproblems and an algorithm for rule-based top-down route derivation. A hierarchical, rule-based approach is aimed at granular planning for generating robust, multi-variant plans.
• Support Vector Machines with Composite Kernels for NonLinear systems Identification
  175 Support Vector Machines, composite kernels, nonlinear system identfication. Amina EL Gonnouni, Abdelouahid Lyhyaoui, Soufiane El Jelali, Manel Martínez Ramón, pages 113 – 118. Show abstract In this paper, a nonlinear system identification based on support vector machines (SVM) has been addressed. A family of SVM-ARMA models is presented in order to integrate the input and the output in the reproducing kernel Hilbert space (RKHS). The performances of the different SVM-ARMA formulations for system identification are illustrated with two systems and compared with the Least Square method.
• The Expert System Approach in Development of Loosely Coupled Software with Use of Domain Specific Language
  113 Expert System, Web Services, Domain Specific Language, Fuzzy Surroundings Piotr Grobelny, pages 119 – 123. Show abstract This paper addresses the problem of supporting the software development process through the artificial intelligence. The expert systems could advise the domain engineer in programming without the detailed experience in programming languages. He will use and integrate, with the help of deductive database and domain knowledge, the previously developed software components to new complex functionalities. The Service Oriented Architecture (SOA) and loosely coupled software allow to fulfill these requirements. The objective of this document is to provide the knowledge representation of atomic Web Services which will be registered as the facts in the deductive database as well as the inferring techniques. Also, the use of Domain Specific Language (DSL) for modeling domain engineer’s requests to the expert system will be considered within this document.
• Coalition formation in multi-agent systemsevolutionary solution
  203 Coalition formation, evolutionary algorithm, multi-agent systems Wojciech Gruszczyk, Halina Kwaśnicka, pages 125 – 130. Show abstract The paper introduces solution of Coalition Formation Problem (CFP) in Multi-Agents Systems (MAS) based on evolutionary algorithm. The main aim of our study is to develop an evolutionary based algorithm for creation of coalitions of agent for solving assumed tasks. We describe the coding schema and genetic operators such as mutation, crossover and selection that occurred to be efficient in solution of CFP. Last part of the document provides a brief comment on our research results.
• Applying Emphasized Soft Targets for Gaussian Mixture Model Based Classification
  92 Classification, estimation, smoothing target, Gaussian Mixture Models Soufiane El Jelali, Abdelouahid Lyhyaoui, Aníbal Ramón Figueiras-Vidal, pages 131 – 136. Show abstract When training machines classifiers, it is possible to replace hard classification targets by their emphasized soft versions so as to reduce the negative effects of using cost functions as approximations to misclassification rates. This emphasis has the same effect as sample editing methods which have proved to be effective for improving classifiers performance. In this paper, we explore the effectiveness of using emphasized soft targets with generative models, such as Gaussian Mixture Models, that offer some advantages with respect to decision (prediction) oriented architectures, such as an easy interpretation and possibilities of dealing with missing values. Simulation results support the usefulness of the proposed approach to get better performance and show a low sensitivity to design parameters selection.
• Improving Naive Bayes Models of Insurance Risk by Unsupervised Classification
  165 classification, cluster analysis, naive bayes, insurance risk Anna Jurek, Danuta Zakrzewska, pages 137 – 144. Show abstract In the paper application of Naïve Bayes model, for evaluation of the risk connected with life insurance of customers, is considered. Clients are classified into groups of different insurance risk levels. There is proposed to improve the efficiency of classification by using cluster analysis in the preprocessing phase. Experiments showed that, however the percentage of correctly qualified instances is satisfactory in case of Naïve Bayes classification, but the use of cluster analysis and building separate models for different groups of clients improve significantly the accuracy of classification. Finally, there is discussed increasing of efficiency by using cluster validation techniques or tolerance threshold that enables obtaining clusters of very good quality.
• On a class of defuzzification functionals
  202 convex fuzzy numbers, ordered fuzzy numbers, defuzzification functional Witold Kosiński, Wiesław Piasecki, pages 145 – 150. Show abstract Classical convex fuzzy numbers have many disadvantages. The main one is that every operation on this type of fuzzy numbers induces the growing fuzziness level. Another drawback is that the arithmetic operations defined for them are not complementary, for instance: addition and subtraction. Therefore the first author (W. K.) with his coworkers has proposed the extended model called ordered fuzzy numbers (OFN). The new model overcomes the above mentioned drawbacks and at the same time has the algebra of crisp (non-fuzzy) numbers inside. Ordered fuzzy numbers make possible to utilize the fuzzy arithmetic and to construct the Abelian group of fuzzy numbers and then an algebra. Moreover, in turn out, that four main operations introduced are very suitable for their algorithmisation. The new attitudes demand new defuzzification operators. In the linear case they are described by the well--know functional representation theorem valid in function Banach spaces. The case of nonlinear functionals is more complex, however, it is possible to prove general, uniform approximation formula for nonlinear and continuous functionals in the Banach space of OFN. Counterparts of defuzzification functionals known in the Mamdani approach are also presented, some numerical experimental results are given and conclusions for further research are drawn.
• Properties of the World Health Organization's Quality of Life Index
  201 Quality of life, inconsistency analysis, consistency-driven pairwise comparisons Tamar Kakiashvili, Waldemar W. Koczkodaj, Phyllis Montgomery, Kalpdrum Passi, Ryszard Tadeusiewicz, pages 151 – 154. Show abstract This methodological study demonstrates how to strengthen the commonly used World Health Organization's Quality of Life Index (WHOQOL) by using the consistency-driven pairwise comparisons (CDPC) method. From a conceptual view, there is little doubt that all 26 items have exactly equal importance or contribution to assessing quality of life. Computing new weights for all individual items, however, would be a step forward since it seems reasonable to assume that all individual questions have equal contribution to the measure of quality of life. The findings indicate that incorporating differences of importance of individual questions into the model is essential enhancement of the instrument.
• Accuracy Boosting Induction of Fuzzy Rules with Artificial Immune Systems
  195 fuzzy logic, artificial immune system, data mining Adam Kalina, Edward Mężyk, Olgierd Unold, pages 155 – 159. Show abstract The paper introduces accuracy boosting extension to a novel induction of fuzzy rules from raw data using Artificial Immune System methods. Accuracy boosting relies on fuzzy partition learning. The modified algorithm was experimentally proved to be more accurate for all learning sets containing non-crisp attributes.
• On Classifcation Tools for Genetic Algorithms
  211 genetic algorithm, entropy, fractal dimension, box-counting dimension Stefan Kotowski, Witold Kosiński, Zbigniew Michalewicz, Piotr Synak, Łukasz Brocki, pages 161 – 164. Show abstract Some tools to measure convergence properties of genetic algorithms are introduced. A classification procedure is proposed for genetic algorithms based on a conjecture: the entropy and the fractal dimension of trajectories produced by them are quantities that characterize the classes of the algorithms. The role of these quantities as invariants of the algorithm classes is discussed together with the compression ratio of points of the genetic algorithm.
• Evolution of Strategy Selection in Games by Means of Natural Selection
  160 Evolutionary Dynamics, Natural Selection, Game Theory, Multi-agent Systems, Agent Design, JADE Daniel Laszlo Kovacs, Tadeus Dobrowiecki, pages 165 – 172. Show abstract In this paper we present a novel agent-based simu­lation model for the natural selection of replicating agents whose survival depends on their programs for selecting their strategy to interact with each other. Game theoretic models de­scribe this kind of interaction. The simulator can be used both for analysis, design and verification of autonomous systems (by intuitively abstracting them into our model and running the simulation). Good system design can be selected even for diffi­cult, underspecified design problems. Although the inspiration of the model comes from evolutionary game theory, the evolving agents may represent not only biological, but also technical sys­tems (e.g. software agents), and thus in some cases the underly­ing evolutionary mechanisms and observations differ from the typically published results.
• Image Similarity Detection in Large Visual Data Bases
  209 image similarity detection, visual databases, morphological spectra, cliques finding in graphs Juliusz Kulikowski, pages 173 – 180. Show abstract A method of similarity clusters detection in large visual databases is described in this work. Similarity clusters have been defined on the basis of a general concept of similarity measure. The method is based also on the properties of mor­phological spectra as a tool for image presentation. In the pro­posed method similarity of selected spectral components in se­lected basic windows are used to similarity of images evalua­tion. Similarity clusters are detected in an iterative process in which non-perspective subsets of images are step-by-step re­moved from considerations. In the method similarity graphs and hyper-graphs also play an auxiliary role. The method is il­lustrated by an example of a collection of medical images in which similarity clusters have been detected.
• Automatic Acquisition of Wordnet Relations by the Morpho-Syntactic Patterns Extracted from the Corpora in Polish
  189 Roman Kurc, Maciej Piasecki, pages 181 – 188. Show abstract In the paper we present an adaptation of the Espresso algorithm of the extraction of lexical semantic relation to specific requirements of Polish. The introduced changes are of more technical character like the adaptation to the existing Polish language tools, but also we investigate the structure of the patterns that takes into account specific features of Polish as an inflectional language. A new method of the reliability measure computation is proposed. The modified version of the algorithm called Estratto was compared with the more direct reimplementation of Espresso on several corpora of Polish. We tested the influence of different algorithm parameters and different corpora on the received results.
• Using UML State Diagrams for Visual Modeling of Business Rules
  96 UML, Rule Systems, Rules Representation, Rules Modeling, Hekate Konrad Kułakowski, Grzegorz J. Nalepa, pages 189 – 194. Show abstract Recently, in response to growing market demand, several different techniques of business rules representation have been created. Some of them try to present business rules in a visual manner. However, due to the complexity of the problem, the graphic representations that are proposed seem to be far from perfection. In this paper we would like to describe how UML state diagrams might be used for business rules formulation and visual modeling. The strength of this approach relies on reusing classical notions provided by UML 2.0, e.g. an action, guard, etc., in a way which is close to theirs original meaning.
• Discovering Technical Analysis Patterns
  205 neural network, time series,data sequence discovery, financial analysis Urszula Markowska-Kaczmar, Maciej Dziedzic, pages 195 – 200. Show abstract In this paper our method of discovering data se­quences in the time series is presented. Two major approaches to this topic are considered. The first one, when we need to judge whether a given series is similar to any of the known pat­terns and the second one when there is a necessity to find how many times within long series a defined pattern occurs. In both cases the main problem is to recognize pattern occurrence(s), but the distinction is essential because of the time frame within which identification process is carried on. The proposed method is based on the usage of multilayered feed-forward neural net­work. Effectiveness of the method is tested in the domain of fi­nancial analysis but its adaptation to almost any kind of se­quences data can be done easily.
• A Hybrid Differential Evolution Algorithm to Solve a Real-World Frequency Assignment Problem
  40 Differential Evolution Algorithm, Frequency Assignment Problem, Evolutionary Computing Marisa Maximiano, Miguel A. Vega-Rodríguez, Juan A. Gómez-Pulido, Juan M. Sánchez-Pérez, pages 201 – 205. Show abstract The Frequency Assignment is a very important task in the planning of the GSM networks, and it still continues to be a critical task for current (and future) mobile communication operators. In this work we present a hybrid Differential Evolution (DE) algorithm to solve a real-world instance of the Frequency Assignment problem. We present a detailed explanation about the hybridization method applied to DE in order to make it more efficient. The results that are shown use accurate interference information. That information was also adopted by other researchers and it represents a real GSM network, granting, therefore, an extremely important applicability. Furthermore, we have analyzed and compared our approach with other algorithms, obtaining good results.
• Hierarchical Rule Design with HaDEs the HeKatE Toolchain
  137 rules, design prototyping, prolog Grzegorz Nalepa, Igor Wojnicki, pages 207 – 214. Show abstract A hierarchical approach to decision rule design in the HeKatE project is introduced in the paper. Within this approach a rule prototyping method ARD+ is presented. It offers a high-level hierarchical prototyping of decision rules. The method allows to identify all properties of the system being designed, and using the algorithm presented in this paper, it allows to automatically build rule prototypes. A practical implementation of the prototyping algorithm is discussed and a design example is also presented. Using the rule prototypes actual rules can be designed in the logical design stage. The design process itself is supported by HaDEs, a set of of design tools developed within the project.
• On Automation of Brain CT Image Analysis
  200 Brain CT Image Analysis, Image Classification, Automatic Image Annotation Mariusz Paradowski, Halina Kwasnicka, Martin Tabakov, Jacek Filarski, Marek Sasiadek, pages 215 – 220. Show abstract During recent years a number of medical diagnosis support systems have been presented. Such systems are important from medical point of view, because they aid physicians in their daily work. Some of those systems, like Computed tomography support systems (CT) rely on image data analysis, making it an interesting research topic from pattern recognition point of view. The paper presents a practical realization of a medical support system, including segmentation, feature extraction and decision processes. Various techniques are used within the system: fuzzy logic, shape descriptors, classifiers and automatic image annotation framework. Series of 2D CT images are an input of the system. As a result, a patient's diagnosis is presented.
• Neuronal Groups and Interrelations
  216 neural networks, pattern recognition, structural pattern recognition, pattern classification Michal Pryczek, pages 221 – 227. Show abstract Recent development of various domains of Artificial Intelligence including Information Retrieval and Text/Image Understanding created demand on new, sophisticated, contextual methods for data analysis. This article formulates Neuronal Group and Extended Neuron Somatic concepts that can be vastly used in creating such methods. Neural interrelations are described using graphs, construction of which is done in parallel with neural network learning. Prototype technique based on Growing Neural Gas is also presented to give more detailed view.
• Semantic Relevance Measure between Resources based on a Graph Structure
  215 Semantic Relevance, Knowledge Management, Information Retrieval, Ontology Sang Keun Rhee, Jihye Lee, Myon-Woong Park, pages 229 – 236. Show abstract Upon developing an information system, establishing an effective information retrieval method is the main problem along with the management of the information. With ontologies, the semantic structure of knowledge can be represented and the resources can be managed along with their context information. Observing that within this knowledge space, resources are not isolated but connected to each other by various types of relations, we believe that utilising those relations can aid information provisioning. In this paper, we explore the concept of the semantic relevance between resources and the semantic structure of a knowledge space, and present a relevance measuring algorithm between resources based on a graph structure, with two implementation examples.
• Exclusion Rule-based Systems—case study
  25 Exclusion Rule-based Systems, decision support systems, railway traffic management system Marcin Szpyrka, pages 237 – 242. Show abstract The exclusion rule-based system , proposed in the paper, is an alternative method of designing a rule-based system for an expert or a control system. The starting point of the approach is the set of all possible decisions the considered system can generate. Then, on the basis of the exclusion rules, the set of decisions is limited to the ones permissible in the current state. A railway traffic management system case study is used in the paper to demonstrate the advantages of the new approach. To compare the exclusion rule-based systems with the decision tables both solutions are considered.
• Ontological Matchmaking in a Duty Trip Support Application in a Virtual Organization
  220 admin Michal Szymczak, Grzegorz Frackowiak, Maria Ganzha, Marcin Paprzycki, Sang Keun Rhee, Jihye Lee, pages 243 – 250. Show abstract In our work, an agent-based system supporting workers in fulfilling their roles in a virtual organization has as its centerpiece ontologically demarcated data. In such system, ontological matchmaking is one of key functionalities in provisioning of personalized information. Here, we discuss how matchmaking will be facilitated in a Duty Trip Support application. Due to the nature of the application, particular attention is paid to the geospatial data processing.
• Computer-aided Detecting of Early Strokes and its Evaluation on the Base of CT Images
  199 ischemic stroke, Ct image, graphics Grzegorz Terlikowski, Elżbieta Hudyma, pages 251 – 254. Show abstract This paper presents an easy way of finding strokes on computer tomography images. By calculating a cohesive rate (CR) of suspicious pixels on a series of CT images there is a possibility of calculating a general probability of a stroke. In a difficult case there is always an opportunity to generate a graph of all stroke probabilities interposed on the original image. It is a very helpful tool for specialists and neurologists working in emergency situations . Supported by grant N518 022 31/1338 (Ministry of Science).

Computer Aspects of Numerical Algorithms

• Numerical Algorithm of Solving the Forced sin-Gordon Equation
  115 numerical algorithm sin-Gordon differential equations riemann function Alexandre Bezen, pages 257 – 261. Show abstract The numerical method of solving the problem of small perturbations of a stationary traveling solution (soliton) of well-known in physics sin-Gordon equation is presented. The solution is reduced to solving a set of linear hyperbolic partial differential equations. The Riemann function method is used to find a solution of a linear PDE. The value of the Riemann function at any particular point is found as a solution of an ordinary differential equation. An algorithm of calculation of a double integral over a triangular integration area is given.
• Merging Jacobi and Gauss-Seidel Methods for Solving Markov Chains on Computer Clusters
  204 Markov chains, parallel algorithms, computer clusters, queuing models, iterative methods, linear equations Jaroslaw Bylina, Beata Bylina, pages 263 – 268. Show abstract The authors consider the use of the parallel iterative methods for solving large sparse linear equation systems resulting from Markov chains—on a computer cluster. A combination of Jacobi and Gauss-Seidel iterative methods is examined in a parallel version. Some results of experiments for sparse systems with over 3× 10^7 equations and about 2× 10^8 nonzeros which we obtained from a Markovian model of a congestion control mechanism are reported.
• Influence of Molecular Models of Water on Computer Simulations of Water Nanoflows
  214 Molecular Dynamics Simulation, molecular models of material, microchannels. Computer Simulation Janusz Bytnar, Anna Kucaba-Piętal, Zbigniew Walenta, pages 269 – 275. Show abstract We present some problems related to the influence of molecular models on Molecular Dynamics simulation of water nanoflows. Out of large number of existing models of water we present some results of the MD simulation of water nanoflows for four molecular models: TIP4P, PPC TIP4P-2005, TIP5P.
• A Sparse Shared-Memory Multifrontal Solver in SCAD Software
  47 sparse direct solver, multi-core computers, finite element method, block factorization Sergiy Fialko, pages 277 – 283. Show abstract A block-based sparse direct finite element solver for commonly used multi-core and shared-memory multiproces­sor computers has been developed. It is intended to solve linear equation sets with sparse symmetrical matrices which appear in problems of structural and solid mechanics. A step-by-step as­sembling procedure together with simultaneous elimination of fully assembled equations distinguishes this method from the well-known multifrontal solver, so this approach can be inter­preted as a generalization of the conventional frontal solver for an arbitrary reordering. This solver is implemented in the com­mercial finite element software SCAD (www.scadsoft.com), and its usage by numerous users has confirmed its efficiency and re­liability.
• Testing Tesla Architecture for Scientific Computing: the Performance of Matrix-Vector Product.
  212 GPU, tesla architecture, BLAS, matrix-vector product, CUDA Paweł Macioł, Krzysztof Banaś, pages 285 – 291. Show abstract The paper presents results of several experiments evaluating the performance of NVIDIA processors, implementing a new Tesla architecture, in matrix-vector multiplication. Three matrix forms, dense, banded and sparse, are considered together with three hardware platforms: NVIDIA Tesla C870 computing board, NVIDIA GeForce 8800 GTX graphics card and one of the newest Intel Xeon processors, E5462, with 1.6 GHz front side bus speed. The conclusions from experiments indicate what speed-ups can be expected when, instead of standard CPUs, accelerators in the form of presented GPUs are used for considered computational kernels.
• Solving a Kind of BVP for Second-Order ODEs Using Novel Data Formats for Dense Matrices
  213 bvp, cache memory, data structures, multicore Przemyslaw Stpiczynski, pages 293 – 296. Show abstract The aim of this paper is to show that a kind of boundary value problem for second-order ordinary differential equations which reduces to the problem of solving tridiagonal systems of linear equations can be efficiently solved on modern multicore computer architectures. A new method for solving such tridiagonal systems of linear equations, based on recently developed algorithms for solving linear recurrence systems with constant coefficients, can be easily vectorized and parallelized. Further improvements can be achieved when novel data formats for dense matrices are used.
• On the hidden discrete logarithm for some polynomial stream ciphers
  169 numerical encryption algorithms, private key, finite automata, implementation Vasyl Ustimenko, pages 297 – 301. Show abstract The paper is devoted to the special key management algorithm for the stream ciphers defined in [12] via finite automata corresponding to the family of directed algebraic graphs of high girth and two affine transformation. Security of the key based on the complexity of the discrete logarithm problem. It has additional heuristic security because of the “hidden base” and “hidden value” of the discrete logarithm function. We consider the heuristic evaluation of the complexity for different attack by the adversary on our private key cipher. The detailed description of the whole algorithm is given. Implemented software package has been used for the evaluation of mixing properties and speed of the private key encryption.
• Empirically Tuning LAPACK's Blocking Factor for Increased Performance
  60 empirical tuning, LAPACK, ATLAS, BLAS, linear algebra Clint Whaley, pages 303 – 310. Show abstract LAPACK (Linear Algebra PACKage) is a statically cache-blocked library, where the blocking factor ( NB) is determined by the service routine ILAENV. Users are encouraged to tune NB to maximize performance on their platform/BLAS (the BLAS are LAPACK's computational engine), but in practice very few users do so (both because it is hard, and because its importance is not widely understood). In this paper we (1) Discuss our empirical tuning framework for discovering good NB settings, (2) quantify the performance boost that tuning NB can achieve on several LAPACK routines across multiple architectures and BLAS implementations, (3) compare the best performance of LAPACK's statically blocked routines against state of the art recursively blocked routines, and vendor-optimized LAPACK implementations, to see how much performance loss is mandated by LAPACK's present static blocking strategy, and finally (4) use results to determine how best to block nonsquare matrices once good square blocking factors are discovered.

Computational Linguistics—Applications

• The incidence of corpus quality and type on topic based text segmentation evaluation: Should we change our evaluation methods in NLP?
  31 Alexandre Labadié, Violaine Prince, pages 313 – 319. Show abstract In this paper, we try to fathom the real impact of corpus quality on methods performances and their evaluations. The considered task is topic-based text segmentation, and two highly different unsupervised algorithms are compared: C99, a word-based system, augmented with LSA, and Transeg, a sentence-based system. Two main characteristics of corpora have been investigated: Data quality (clean vs raw corpora), corpora manipulation (natural vs artificial data sets). The corpus size has also been subject to variation, and experiments related in this paper have shown that corpora characteristics highly impact recall and precision values for both algorithms.
• N-gram language models for Polish language. Basic concepts and applications in automatic speech recognition systems.
  219 language models, language modeling, POS-tagging, ASR, Polish language, n-gram models Bartosz Rapp, pages 321 – 324. Show abstract Usage of language models in automatic speech recognition systems usually give significant quality and certainty improvement of recognition outcomes. On the other hand, wrongly chosen or trained language models can result in serious degradation not only recognition quality but also overall performance of the system. Proper selection of language material, system parameters and representation of the model itself is important task during language models construction process. This paper describes basic aspects of building, evaluating and applying language models for Polish language in automatic speech recognition systems, which are intended to be used by lawyer's chambers, judiciary and law enforcements. Language modeling is a part of project which is still early stage of development and work is ongoing so only some basic concepts and ideas are presented in this paper.
• SemCAPTCHA―user-friendly alternative for OCR-based CAPTCHA systems
  183 CAPTCHA, positive semantic priming, linguistic competence, OCR Paweł Łupkowski, Mariusz Urbański, pages 325 – 329. Show abstract In this paper we present a new CAPTCHA system (Completely Automated Turing Test To Tell Computers and Humans Apart ). This proposal, SemCAPTCHA, is motivated by an increasing number of broken OCR-based CAPTCHA systems and it is based not only on text recognition but also on text understanding. We describe SemCAPTCHA from both user's perspective and system's perspective and compare it to some currently popular CAPTCHAs. We also briefly describe an experiment carried out to test our CAPTCHA on human users.
• Arabic/English Word Translation Disambiguation Approach based on Naive Bayesian Classifier
  145 parallel corpus, word translation disambiguation, Arabic language, cross language information retrieval Farag Ahmed, Andreas Nürnberger, pages 331 – 338. Show abstract We present a word sense disambiguation approach with application in machine translation from Arabic to English. The approach consists of two main steps: First, a natural language processing method that deals with the rich morphology of Arabic language and second, the translation including word sense disambiguation. The main innovative features of this approach are the adaptation of the Naïve Bayesian approach with new features to consider the Arabic language properties and the exploitation of a large parallel corpus to find the correct sense based on its cohesion with words in the training corpus. We expect that the resulting system will overcome the problem of the absence of the vowel signs, which is the main reason for the translation ambiguity between Arabic and other languages.
• Smarty―Extendable Framework for Bilingual and Multilingual Comprehension Assistants
  107 comprehension assistant, WordNet, BalkaNet, intelligent dictionary, Bulgarian, English, lexical resources, word-sense disambiguation Todor Arnaudov, Ruslan Mitkov, pages 339 – 344. Show abstract This paper discusses a framework for development of bilingual and multilingual comprehension assistants and presents a prototype implementation of an English-Bulgarian comprehension assistant. The framework is based on the appli­cation of advanced graphical user interface techniques, Word­Net and compatible lexical databases as well as a series of NLP preprocessing tasks, including POS-tagging, lemmatisation, multiword expressions recognition and word sense disambigua­tion. The aim of this framework is to speed up the process of dictionary look-up, to offer enhanced look-up functionalities and to perform a context-sensitive narrowing-down of the set of translation alternatives proposed to the user.
• SuperMatrix: a General Tool for Lexical Semantic Knowledge Acquisition
  163 Bartosz Broda, Maciej Piasecki, pages 345 – 352. Show abstract The paper presents the SuperMatrix system, which was designed as a general tool supporting automatic acquisition of lexical semantic relations from corpora. The construction of the system is discussed, but also examples of different applications showing the potential of SuperMatrix are given. The core of the system is construction of co-incidence matrices from corpora written in any natural language as the system works on UTF-8 encoding and possesses modular construction. SuperMatrix follows the general scheme of distributional methods. Many different matrix transformations and similarity computation methods were implemented in the system. As a result the majority of existing Measures of Semantic Relatedness were re-implemented in the system. The system supports also evaluation of the extracted measures by the tests originating from the idea of the WordNet Based Synonymy Test. In the case of Polish, SuperMatrix includes the implementation of the language of lexico-syntactic constraints delivering means for a kind of shallow syntactic processing. SuperMatrix processes also multiword expressions as lexical units being described and elements of the description. Processing can be distributed as a number of matrix operations were implemented. The system serves huge matrices.
• Definition Extraction: Improving Balanced Random Forests
  154 definition extraction, machine learning, balanced random forests Łukasz Degórski, Łukasz Kobyliński, Adam Przepiórkowski, pages 353 – 357. Show abstract The article discusses methods of improving the ways of applying Balanced Random Forests (BRFs), a machine learning classification algorithm, used to extract definitions from written texts. These methods include different approaches to selecting attributes, optimising the classifier prediction threshold for the task of definition extraction and initial filtering by a very simple grammar.
• An application supporting language analysis within the framework of the phonetic grammar
  120 language analysis, phonetics, phonetic grammar, data-mining, distances Krzysztof Dyczkowski, Norbert Kordek, Paweł Nowakowski, Krzysztof Stroński, pages 359 – 362. Show abstract The aim of the paper is to present an application supporting language analysis within the framework of the phonetic grammar. The notion of the phonetic grammar has been concisely introduced and the basic potential of the application and the algorithms employed in it are briefly discussed. The application is to enable a uniform description and a comparative analysis of many languages. At the first stage the languages taken into consideration are Polish, Chinese and Hindi.
• Set of active suffix chains and its role in development of the MT system for Azerbaijani
  61 Machine translation, formal parsing, Azerbaijani, Turkic languages group Rauf Fatullayev, Ali Abbasov, Abulfat Fatullayev, pages 363 – 368. Show abstract Set of active suffix chains and its role in development of the MT system for AzerbaijaniAbstract—Definition process of the active suffix chains of Azerbaijani (The Azerbaijani language) has been explained in this paper. Morphologically Azerbaijani word-forms consist of the stem and simple or compound suffixes (suffix chains). Be­cause Azerbaijani is an agglutinative language the simple suf­fixes in this language can form the great number of suffix chains and consequently, it is possible to generate practically “endless” number of word-forms from the same stem. While de­veloping machine translation system from Azerbaijani into non-Turkic languages for the translation of any word-form it is nec­essary to translate its suffix chain in whole. It is not possible to get the correct transla­tion of a word-form by translating each simple suffix of its suffix chain separately and then by putting these transla­tions together. For these reasons, it is necessary to define the subset of suffix chains frequently used in the texts in­stead of the set of every possible suffix chains.
• A New Word Sense Similarity Measure in WordNet
  75 Semantic Similarity,Information Content,Lexical Database,WordNet Ali Sebti, Ahmad Abodollahzadeh Barfroush, pages 369 – 373. Show abstract Recognizing similarities between words is a basic element of computational linguistics and artificial intelligence applications. This paper presents a new approach for measuring semantic similarity between words via concepts. Our proposed measure is a hybrid system based on using a new Information content metric and edge counting-based tuning function. In proposed system, hierarchical structure is used to present information content instead of text corpus and our result will be improved by edge counting-based tuning function. The result of the system is evaluated against human similarity ratings demonstration and shows significant improvement in compare with traditional similarity measures.
• A Linguistic Light Approach to Multilingualism in Lexical Layers for Ontologies
  155 Alexander Troussov, John Judge, Mikhail Sogrin, Amine Akrout, Brian Davis, pages 375 – 379. Show abstract Semantic web ontologies are being increasingly used in modern text analytics applications and Ontology-Based Information Extraction(OBIE) as a means to provide a semantic backbone either for modelling the internal conceptual data structures of the Text Analytics(TA) engine or to model the Knowledge base to drive the analysis of unstructured information in raw text and subsequent Knowledge acquisition and population. Creating and targeting Language Resources(LR)s from a TA to an Ontology can be time consuming and costly. In \cite{lrecpaper} the authors describe a user-friendly method for Ontology engineers to augment an ontologies with a lexical layer which provides a flexible framework to identify term mentions of ontology concepts in raw text. In this paper we explore multilinguality in these lexical layers using the same framework. We discuss a number of potential issues for the “linguistic light” Lexical Extensions for Ontologies (LEON) approach when looking at languages more morphologically rich and which have more complex linguistic constraints than English. We show how the LEON approach can cope with these phenomena once the morphological normaliser used in the lexical analysis process is able to generalise sufficiently well for the language concerned.
• Modeling the Frequency of Phrasal Verbs with Search Engines
  97 phrasal verbs, search engines, quantitative language modeling Dawid Weiss, Grażyna Chamielec, pages 381 – 388. Show abstract There are well over a thousand phrasal verbs in English. For non-native speakers they are notoriously difficult to remember and use in the right context. We tried to construct a ranking of phrasal verbs according to their estimated occurrence frequency, based on quantitative information available from the public indexable Web. Technically, we used major Web search engines to acquire phrase-occurrence statistics, measured consistency between the rankings implied by their results and confirmed that a rough set of `classes' of phrasal verbs can be distinguished. While this technique relies on inaccurate and possibly biased estimation functions, we show that the overall distribution of ranks seems to be consistent among all the queried search engines operated by different vendors.

8th International Multidisciplinary Conference on e-Commerce and e-Government

• On stimulus for citizens' use of e-government services
  46 e-governmemnt, motivation for use, EU27 Cene Bavec,