Please follow this link to the new website of the Institute of Computer Science ...
+ -
Duration 
2019-01-01 -2022-12-30Introduction of targeted protection of cereal crops against insect pests in precision farming
Number:QK1910281
Investigator:
Co-investigator:
Anotation:
This project will develop statistical tools for prediction of crop pests in the age of precision agriculture. The developed methodologz will be based on modern semiparametric and dznamical modeling in the GAM framework. The models will be developed in several variants and the most suitable model will be selected by formalized statistical procedures. Based on the validated model, we will construct both routine predictions and derive recommendations for crop management timing.
2018-05-01 -2020-04-30Enhancing human resources for research in theoretical computer science
Number:CZ.02.2.69/0.0/0.0/17_050/0008361
Investigator:Mgr. Igor Sedlár Ph.D.
Co-investigator:
Anotation:
Theoretical computer science uses mathematical methods in order to clarify some of the most important notions used in computer science and, therefore, it is a key research area within computer science. Logic is an important part of theoretical computer science, focusing on modelling bodies of information and ways how agents use them in reasoning. So called non-classical logics, alternatives to classical logic, aim at a better representation of various aspects of information and reasoning than the representations based on classical logic (these aspects include vagueness, incompleteness and inconsistency of information or limitations of reasoning agents). Within this project, the Institute of Computer Science hosts three researchers working on non-classical logics and their applications to specific problems in theoretical computer science. The overall goal of the project is to study (i) applications of non-classical logics in program verification, (ii) applications of non-classical logics in the study of weighted structures (with focus on the Valued Constraint Satisfaction Problem) and (iii) non-classical logics with so-called generalized quantifiers (with focus on their computational properties). This project is funded by the European structural and investment funds within the Operational Programme Research, Development, Education.
2018-01-01 -2020-06-30Urbanization of weather and air quality forecast and climatic scenarios for Prague
Number:CZ.07.1.02/0,0/0,0/16_040/0000383
Investigator:
Co-investigator:
Anotation:
The project Urbi Pragensi addresses improvements and implementation of the weather prediction and air quality prediction for region of Prague together with more detailed assessment of impacts of climate change in the city. The prediction improvements are achieved by utilization of the modern methods based on incorporation of parameterization of urban level processes into the high resolution atmospheric models. Complementary approach are microscale simulations which can expose in detail situation in given parts of the town with substantial burden of heat island and air pollution. This can contribute to estimation of the health risks as well as to efficiency assessment of proposed mitigation measures. ICS is involved in works on concept KK1 (meteorological prediction), KK2 (air quality prediction) and it is the coordinator of the concept KK4 (microscale simulations). The main partner of the project is Charles University in Prague, another partner is Czech Hydrometeorological Institute in Prague. In the scope of the project, we tightly collaborate with other Czech and foreign academic institutions (from Germany and Finland). We also tightly cooperate with principal intended users of the project outputs (e.g. Prague municipality and Prague Institute of Planning and Development).
2017-04-01 -2020-03-31Collaboration on Fermilab experiments
Number:CZ.02.1.01/0.0/0.0/16_013/0001787
Investigator:
Co-investigator:
Anotation:
The subject of the project is to improve the methods for data processing of Fermilab experiments and design of new analyzing procedures of these data based on machine learning algorithms and artificial intelligence. The goal is to continually improve the potential of acquiring new scientific knowledge on collaborative experiments in Fermilab, the development of new innovative data processing methods in high energy physics and the improvements of computing infrastructures of participating experimental institutions in Fermilab. The target group of the project results are the researchers of universities and research and development experts in general. From project outputs benefit experts from all research institutions cooperating experiments in Fermilab.
2017-01-01 -2017-12-31zkouska
Number:0001
Investigator: RNDr. Ladislav Andrej CSc.
Co-investigator:
Anotation:
2017-01-01 -2019-12-31Metalearning for Extraction of Rules with Numerical Consequents
Number:GA17-01251S
Investigator:prof. Ing. RNDr. Martin Holeňa CSc.
Co-investigator:
Anotation:
One of the key directions of data mining, particularly important if human-comprehensibility plays a role, is the extraction of rules from data. The project aims at rules with consequents corresponding to numerical variables. In spite of the ubiquity of such variables, rules extraction is not yet as mature for them as for classification and association rules. The main objective of the project is to develop a framework to enable assessing different algorithms for the extraction of rules with numerical consequents from a given dataset. Traditional algorithms view consequent variables as responses and antecedent variables as regressors of regression models. They are complemented by emerging algorithms of computational topology. The framework will be based on metalearning, i.e., learning from metadata about the past performance of the algorithms on datasets with similar values of metafeatures. Metalearning has been for several decades successfully used in classification and some other areas of data mining, but its application to the extraction of this kind of rules is novel. Targets: development of a metalearning framework for the extraction of rules with numerice consequents, search for metafeatures in two selected application domains, research into robust metalearning, validation of the metalearning framework on metadata from both application domains.
2017-01-01 -2019-12-31Predicate graded logics and their applications to computer science
Number:GA17-04630S
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
Classical mathematical logic, built on the conceptually simple core of propositional Boolean calculus, plays a crucial role in modern computer science. A critical limit to its applicability is the underlying bivalent principle that forces all propositions to be either true or false. Propositional logics of graded notions (such as tall, rich, etc.) have been deeply studied for over two decades but their predicate extensions (accommodating, among others, modalities and quantifiers) are still only very partially developed and scarcely applied to particular komputer science problems. The overall goal of the proposed project is to develop predicate graded logics in two complementary directions: (1) studying logical systems in full generality in order to provide a solid mathematical framework and (2) applying achieved results to three particular problems in computer science which heavily involve graded notions: representation of vague and uncertain knowledge, valued constraint satisfaction problems, and modelling of coalition games. We plan to develop predicate graded logics by giving them solid mathematical foundations and applying the achieved results to three particular computer science problems involving graded notions: management of uncertainty, valued constraint satisfaction problems, and modelling of coalition games.
2017-01-01 -2019-12-31Strength of materials and mechanical components based on iron: Multi-scale approach
Number:GA17-12925S
Investigator:
Co-investigator:doc. Dr. Ing. Miroslav Rozložník
Anotation:
Ductile or brittle behavior of cracks is one of the key phenomena which may have a crucial influence on static and dynamic strength of mechanical structures utilizing bcc iron based materials, e.g. ferritic steels. Continuum predictions on ductile/brittle behavior of a central crack under biaxial tension show that the change of called T-stress can change ductile crack behavior to brittle crack extension. We utilize 3D atomistic molecular dynamic (MD) simulations in bcc iron at various temperatures to verify predictions on ductile-brittle transition caused by T-stress. It will be done for central cracked specimens under biaxial tension and as well for edge cracked samples under uniaxial tension, available for experiments. The topic is important for reactor pressure vessels and interpretation of fracture experiments. Another important aim is interconnecting with first-principles calculations for model clusters of restricted size, pointed at cohesive energy, tension and shear strength, atomic configurations and forces at defects, determining interatomistic potential parameters for MD.
2016-03-01 -2019-02-28SYSMICS: Syntax meets semantics: Methods, interactions, and connections in substructural logics
Number:689176
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
Substructural logics are formal reasoning systems that refine classical logic by weakening the structural rules in Gentzen sequent calculus. While classical logic formalises the notion of bivalent truth, substructural logics allow to handle notions such as resources, partial truth, meaning, and natural language syntax, motivated by studies in computer science, epistemology, economy, and linguistics. Traditionally, substructural logics have been investigated following three main approaches: proof theoretic, algebraic and abstract-algebraic. Although some connections among these approaches were observed long ago, in large part these practices developed in independence. The main objective of this project is to establish a network of leading experts from these three areas, with the aim of reuniting these traditions and their communities and obtain new deep results.
2016-01-01 -2018-12-31Extremal graph theory and applications
Number:16-07822Y
Investigator:Mgr. Diana Piguet Ph.D.
Co-investigator:
Anotation:
Abstrakt anglicky Graphs are among the simplest mathematical structures. They form the foundation for much of Computer Science and their importance has grown enormously with the development of computer networks. Extremal graph theory focuses on interactions between different properties of graphs. In this project we link extremal graph theory to several other fields, including probability theory, analysis and geometry. We exploit novel techniques that were developed for embedding problems in sparse graphs and those that arose from the theory of dense graph limits. The aim of the project is to develop general tools that relate to the Szemerédi Regularity lemma, the Stability method, extremal problems in graphons, and applications of the Chatterjee-Varadhan approach to large deviations of Erdős-Rényi random graphs. Among our main goals are the resolution of the Loebl-Komlós-Sós conjecture, applications of extremal graph theory in geometric combinatorics, or work on the "infamous upper tail problem" for subgraph counts in random graphs.
2016-01-01 -2019-12-31Research infrastructure for Fermilab experiments
Number:LM2015068
Investigator:
Co-investigator: Ing. František Hakl CSc.
Anotation:
RI serves for Czech contribution to particle physics research on experiments at Fermilab. It consists of experiments on which Czech physicists collaborate in Fermilab and of infrastructures of the Czech collaborating institutions. Members of RI work on the Fermilab's experiments NOvA, D0 and plan to join a new experiment in two years to contribute to its design and construction. In the Czech Republic it is a RCCPP computing farm and physics laboratory in FZU, cluster for artificial intelligence and neural networks algorithms in ICS and numerical and statistical computing servers at CTU. The whole infrastructure serves for particle physics experiments and for researchers for many years. The RI as top world research environment serves also for education of undergraduate and postgraduate students.
2015-10-01 -2017-12-31Spolupráce na experimentech ve Fermiho národní laboratoři, USA
Number:LG15047
Investigator:
Co-investigator: Ing. František Hakl CSc.
Ing. Marcel Jiřina DrSc.
Anotation:
Spolupráce na excelentních projektech ve Fermiho národní laboratoři (Fermilab), Batavia, USA. Fermilab je špičková americká laboratoř pro fyziku částic. Aktuálně spolupracujeme na končícím experimentu D0, který zkoumá srážky protonů s antiprotony. Experiment poslední dvě dekády prezentovat významné výsledky na mezinárodních konferencích a čeští spolupracovníci jsou spoluautoři několika set vědeckých publikací. Dále spolupracujeme na běžícím experimentu NOvA, tzv. "long base line" neutrinovém experimentu, který měří základní parametry oscilací neutrin pomocí dvou 800km od sebe vzdálených detektorů. Připravujeme se na spolupráci na další generaci experimentů ve Fermilab.
2015-01-01 -2017-12-31Automated Knowledge and Plan Modeling for Autonomous Robots
Number:P103-15-19877S
Investigator:
Co-investigator: Mgr. Roman Neruda CSc.
Anotation:
Artificial Intelligence reached maturity in many of its subareas and the most recent trend is re-integration of developed techniques to tackle hard real-life problems such as driverless cars, deep-space and ocean explorers, drones etc. The project deals with developing autonomous agents (robots) that can build and refine their internal models and do their own decisions. The focus is on internal knowledge model of autonomous agents that is appropriate for planning their behavior and that can be obtained and refined in a (semi-) automated way rather than being manually encoded. The model will be a core part of a modular architecture and it will be developed by integrating research results from areas of robotics, planning, uncertainty reasoning, knowledge representation, nature-inspired computation, and linguistics. In addition to theoretical formal results (models, algorithms), the important output will be verification of the developed techniques on real robots. The aim is to bridge different research areas and bring their results closer to practical applicability.
2015-01-01 -2017-12-31Estimation of psychometric measures as part of admission test development
Number:P407-15-15856Y
Investigator:RNDr. Patrícia Martinková Ph.D.
Co-investigator:
Anotation:
Project aims to describe the current status of admission process to Czech colleges and universities and to prepare a methodology for development of standardized admission tests. Project covers the whole cycle of test development – from defining objectives, blueprinting, item writing, reviewing and pretesting, to assembling the test and its validation. Special attention is given to psychometric tools needed for student scoring, item calibration, detection of improper items, estimation of reliability and validity in complex design and student scoring and item selection in automatized adaptive testing. Theoretical results are implemented and demonstrated on real data from admission test to a medical school.
2015-01-01 -2017-12-31Model complexity of neural, radial, and kernel networks
Number:P202-15-18108S
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The project will contribute to development of theoretical foundations of neurocomputing. The goal of the research is to obtain new knowledge in terms of mathematical results describing capabilities and limitations of multilayer networks. Relationships between networks with various types of computational units (perceptrons, radial, and kernel), various parameters of these units will be described. Estimates of model complexities of networks will be derived in dependence on input dimensions, types of units and network architectures. Properties of highdimensional tasks which can be represented or approximated by networks of reasonable sizes will be characterized. Optimal solutions of learning tasks from point of view of generalization and model complexity will be analyzed.
2015-01-01 -2017-12-31Modelling vague quantifiers in mathematical fuzzy logic
Number:P202-I1897-N25
Investigator:doc.Ing. Petr Cintula Ph.D.
Co-investigator:
Anotation:
Vague quantifiers like `many', `few', or `about a half' present a major problem in natural language processing. Designing a satisfactory theory of vague quantifiers requires to construct formal models and evaluate them with regard to linguistic adequateness, automated deduction, and embeddability in logical frameworks. This constitutes a serious research challenge involving computer science, logic, linguistics, and analytic philosophy. The fuzzy logic paradigm, based on the notion of degrees of truth, provides a mathematical apparatus for dealing with several aspects of vagueness. The applications of fuzzy methods to vague quantifiers have so far largely neglected the potential of deductive systems studied by mathematical fuzzy logic. The aim of the project is to deepen and extend the mathematical foundations for adequate modeling of vague quantifiers by employing formalisms and results of mathematical fuzzy logic, including modal logics with two-level syntax, game-theoretic semantics, and automated reasoning techniques.
2015-01-01 -2017-12-31Simulation-Based Computation of Robust Invariants of Hybrid Dynamical Systems
Number:P202-15-14484S
Investigator:doc.Dipl.Ing. Stefan Ratschan Dr.-techn.
Co-investigator:
Anotation:
One of the most important objects used in the field of formal verification are invariants. An invariant is a set of states of a given system such that the system will always stay in this set of states. Recently, there has been a lot of progress on the constraint-based computation of invariants, that reduces invariant computation to a constraint solving problem in a decidable theory. However, this approach cannot be applied in cases where the corresponding logical theory is not decidable, or where the available constraint solvers are not efficient enough to solve problems of interesting size. The proposal concerns computation of invariants of hybrid dynamical systems, that is, dynamical systems that have partially discrete, partially continuous states and behavior. In order to circumvent the mentioned problems of undecidability and efficiency limitations of constraint solvers, we use a radically different approach that exploits robustness and simulations.
2015-01-01 -2017-12-31Totally ordered monoids
Number:P201-15-07724Y
Investigator:Ing. Milan Petrík Ph.D.
Co-investigator:
Anotation:
The subject of the project is an investigation of the variety of lattice-ordered monoids with a special focus on the subclass of totally ordered ones. As the known methods investigate this problem seem to become exhausted, the project intends to take a benefit of newly introduced methods of geometric nature to attack this task, whose nature is algebraic, in a new way. The new tools include the approach of web geometry, a branch of the differential geometry introduced by Blaschke and Bol, and the representation by Cayley monoids; both these approaches allow to display algebraic properties of structures in an appealing visual way.
2014-04-01 -2016-12-31Klimatické sítě: Rozmanitost měřítek dynamiky a interakcí v atmosféře Země
Number:LH14001
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
2014-01-01 -2016-12-31Large-scale dynamics and critical transitions in neuronal networks and their role in limbic seizure genesis
Number:P303/14-0263S
Investigator:
Co-investigator: RNDr. Milan Paluš DrSc.
Anotation:
Transition to epileptic seizure represents sudden and abrupt shift between distinct dynamic regimes of the brain. In simplified brain preparations we have demonstrated that seizures are preceded by detectable changes in neuronal behaviour which marked progressive decrease in neuronal network resilience and proximity to transition to seizure. These processes correspond to phenomenon of “critical slowing down” described in dynamics of complex systems, innovative and rapidly emerging field of modern physics. In the proposed work we aim to elucidate mechanisms responsible for transition to seizure in intact brain and examine whether this process displays features of critical transition. We will apply integrative approach which will combine advanced techniques of large-scale recording, methods of active probing, computational modelling and analyses of complex systems. Demonstrating that dynamics of epileptic networks is governed by similar principles to other dynamical systems will open new ways to design innovative and more efficient therapies to abort or reverse transition to seizure
2013-03-21 -2016-05-31Modeling of Complex Systems by Soft-Computing Methods
Number:LD13002
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is theoretical analysis of properties of softcomputing computational models suitable for processing of high-dimensional complex data from point of view of minimization of model complexity, efectivity of learning and capability of generalization. Further goal is application of theoretical results to design of hybrid algorithms of methalearning with adaptive choices of models and their parameters and implementation of these algorithms as Java and MATLAB software tools and their testing on real data.
2013-02-01 -2016-01-31Constructing Advanced Comprehensible Classifiers
Number:13-17187S
Investigator:prof. Ing. RNDr. Martin Holeňa CSc.
Co-investigator:
Anotation:
The proposed project will further develop key directions of research into methods attempting to alleviate the discrepancy between accurate and comprehensible classifiers. These are the extraction of logical rules from trained neural networks, and the evolution of sets of comprehensible classification rules by means of genetic, ant colony, and similar optimization algorithms. The project aims at increasing the accuracy of rule induction in classification rules mining, and at the elaboration of new methods for inferring comprehensible rules from accurate classifiers, such as support vector machines. It will also perform theoretical research in this area, in particular research into the relationship that the accuracy-comprehensibility trade-off has to the difference between descriptive and generative classifiers, and will search for a suitable formalization of the concept of classification comprehensibility.
2013-02-01 -2016-01-31Personality and spntaneous brain activity during rest and movie watching: relation and structural determinants
Number:13-23940S
Investigator:Ing. Mgr. Jaroslav Hlinka Ph.D.
Co-investigator:
Anotation:
Current psychological theory provides complex description of mental functions and personality. It is generally accepted that mental functions have brain as their substrate, suggesting that personality differences should be reflected in brain structure and function. However, the specific relation remains elusive. A rapidly developing area of brain research is the study of spontaneous brain activity with functional magnetic resonance imaging, allowing simultaneous characterization of a plethora of brain networks. We suggest that psychological characteristics are likely to be strongly related to endogenous patterns of brain activity. We seek to further improve the detectability of this relation by robustifying the patterns by measuring the brain during natural viewing, a rich quasi-realistic stimulation. The combined psychometrical and neuroimaging study will allow relating specific features of functional brain networks and personality. The findings will be put in context by investigation of structural determinants of this relation through data-driven analysis and theoretical models.
2013-02-01 -2017-01-31Advanced random field methods in data assimilation for short-term weather prediction
Number:13-34856S
Investigator:prof. RNDr. Jan Mandel CSc.
Co-investigator:
Anotation:
An indispensable feature of weather nowcasting is data assimilation (DA), which includes the newest data, e.g. radar and satellite data, into the numerical weather prediction model. The goal of the project is to introduce qualitatively new DA techniques. These techniques (automatic registration, morphing) have arisen in the theory of pattern recognition. They perform a correction of shape and position of objects by means of a horizontal motion ?eld at the same time as the correction of the values of the physical variables. The methods will be generalized so as to be able to use radar data. This will result in better timing and position of atmospheric fronts or precipitation ?elds and in an improvement of the forecast. Many DA techniques are computationally demanding. New methods which rely on the theory of random ?elds and Wavelet Transform will be developed. Expensive computations with large matrices in the Ensemble Kalman Filter analysis will be replaced by cheap wavelet transform calls. The methods also can substantially reduce the number of ensemble members required.
2013-02-01 -2017-01-31An Order-Based Approach to Non-Classical Propositional and Predicate Logics
Number:13-14654S
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
Formal systems of (non-)classical logics are essential in many areas of computer science. Their appreciation is due to their deductive nature, universality and portability, and the power they gain from their mathematical background. Such a diverse landscape of logical systems has greatly benefited from a unified approach offered by Abstract Algebraic Logic. The purpose of this project is to develop a variant of this theory, based on the notion of ordered semantics and its interplay with implication connective. We aim at a stronger, better applicable abstract theory for both propositional and predicate logics. As a showcase, we plan to illustrate the power of the resulting theory on two important families of non-classical logics: substructural and fuzzy ones.
2013-02-01 -2018-01-31Iterative Methods in Computational Mathematics: Analysis, Preconditioning, and Applications
Number:13-06684S
Investigator:prof. Ing. Miroslav Tůma CSc.
Co-investigator:
Anotation:
The project deals with iterative methods for several important problems of numerical linear algebra. It includes their analysis, preconditioning, solving ill-posed problems as well as real-world applications. We focus on Krylov subspace methods, open questions related to their convergence, associated matrix approximation problems, error estimation and stopping criteria. We will study various preconditioning techniques including new algorithms based on incomplete factorizations and orthogonalization schemes, and block saddle-point preconditioning. We intend to analyze regularization methods for solving ill-posed problems in image and signal processing, open problems in total least squares and Golub-Kahan bidiagonalization. An inseparable part of our work are broad international collaboration and selected real-world applications such as the approximation of scattering amplitude and nuclear magnetic resonance.
2013-01-01 -2014-12-31Game - theoretical approach to many - valued logics
Number:412
Investigator:RNDr. Ondrej Majer CSc.
Co-investigator:
Anotation:
The goal of the project is to continue the longstanding cooperation between Prague and Vienna on the application of game and theory fuzzy logic. The project will focus particulary on the following research issues: a) the relationship between game semantics for fuzzy logic (dialogical/evaluation games) and their relationship to other game–theoretical many–valued semantics (probalistic semantics for independent-friendly logic) b) consequences of the game semantics for the study of fuzzy logics (e.g., the problem of safe models) c) application of kooperative games in many-valued logics – the formativ of fuzzy coalitions and bargaining d) applicatin of dialogical games in linguistics with a focus on generalized quantifiers The output of the project will consist of papers in etablished international journals.
2012-07-01 -2015-06-30
Number:470
Investigator:prof. Ing. Miroslav Tůma CSc.
Co-investigator:
Anotation:
2012-06-01 -2014-05-31100 vědců do středních škol
Number:CZ.1.07/2.3.00/35.0020
Investigator: Doc. Ing. Mgr. Petr Klán CSc.
Co-investigator:
Anotation:
Cílem projektu je pomoci vědcům ke komunikaci se středoškolskými studenty a pedagogy. Vzájemně se propojit s cílem zlepšit výuku matematiky a obecně vědy na středních školách a vytvořit větší vlnu inovativních studentů se zájmem uskutečnit nápady pomocí samostatných projektů a později se stát vědcem. Dalším cílem je pomoci středoškolským studentům a učitelům motivačně usktečňovat samostatné projekty synergicky s účastí na přehlídce nejlepších projektů maldých středoškolských vědců EUCYS (European Union Contest for Young Scientists) a nejlepším z nich potom na ISEF v USA (International Scientific and Engineering Fair) a účastnit se tam odborného programu.
2012-03-01 -2014-12-31
Number:LG12020
Investigator:
Co-investigator: Ing. František Hakl CSc.
Anotation:
2012-03-01 -2014-12-31
Number:LD12009
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
2012-01-01 -2013-12-31Number theory and its applications
Number:7AMB12SK128
Investigator:
Co-investigator:prof. RNDr. Štefan Porubský DrSc.
Anotation:
Aim of project is to provide a frame for a process of proving new results which could be published in international journals and as wel as presented on international conferences. The graduate students will be also integrated in the project team and the outcome of the project will also be used in the teaching process. The work of the project falls is mathematics, more precisely the number theory. The subject of research will be the distribution properties of number sequences. These properties have besides their theoretical importance also a practical utilizacion, for they are connected with the evaluation of pseudo random number generators. The relevance of such research is growing due to the development of communication technologies (e.g. in bank services, evaluation of security of IT structures, etc.) many of which are utizing cryptographic protocols based on the distribution properties of generated number sequences. Another field of applications are the quasi Monte Carlo numerical methods.
2012-01-01 -2014-12-31
Number:TA02031411
Investigator:
Co-investigator: Ing. Marek Brabec Ph.D.
Anotation:
2012-01-01 -2014-12-31A Multivalued Approach to Optima and Equilibria in Economics
Number:P402/12/1309
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
A multivalued approach offers effecient tools for dealing with game-theoretic models and analysing optima and economic equilibria. The goal of this project is to use the methods based on multivalued analysis and logics for the investigation of several selected challenging problems that arise in game theory. This will include both cooperative and non-cooperative games. We will analyse the stability and sensitivity of generalized Nash equilibrium problems and optimise the construction of the bidding strategy at the electricity spot market of the European Energy Exchange. A particular attention will be paid to solutions of cooperative games with fuzzy coalitions. New dynamic bargaining procedures for the core of such cooperative games will be developed. The applications of many-valued and dynamic epistemic logics will enable us to capture the flow of information and knowledge in games and to model the imprecision of data. The multivaluedness is present in all considered game models both on the side of the game description and the game solutions.
2012-01-01 -2014-12-31Geometry of associative structures
Number:P201/12/P055
Investigator:Ing. Milan Petrík Ph.D.
Co-investigator:
Anotation:
This project intends to investigate associative operations and structures from a geometrical point of view. The main motivation comes from the study of web geometry which is a branch of the di erential geometry. As some results have already shown, the concepts of this discipline can be adopted successfully to characterize the associativity of totally ordered monoids and triangular norms in a very intuitive visual way. Thus, there is a good motivation to continue this research and to exploit its potential. The aim of the project is a geometric investigation of questions of structural characterization of associative structures, in particular, totally ordered monoids which play an important role in MTL algebras. These algebras represent semantics of the monoidal t-norm based logic which is a prototypical many-valued logic studied intensively nowadays by many researches. As the outcome of the project, it is expected not only a deeper understanding of associative structures but also a generalization of web geometry to a more general case.
2012-01-01 -2014-12-31Integrated Verification and Falsification of Hybrid Systems of Industrial Size
Number:P202/12/J060
Investigator:doc.Dipl.Ing. Stefan Ratschan Dr.-techn.
Co-investigator:
Anotation:
A large part of the development costs of complex systems with a deep integration of software and physical components (e.g. cars. trains, airplanes) goes into the stage of testing the correctness of the system design. The notion of a hybrid system is a formalism to model such systems. The results of this project will, for the first time, allow techniques for formal verification of hybrid systems to handle systems of industrial size. This will be achieved by integrating techniques for formal verification and automated falsification/testing. Since new results in hybrid systems falsification allow the handling of huge systems already now, and since, in the case of software, a corresponding integration of verification and falsification has resulted in huge speedups, we are convinced that this approach will result in a corresponding break-through also in the field of hybrid systems.
2012-01-01 -2015-12-31Distribution and metric properties of number sequences and their applications
Number:P201/12/2351
Investigator:prof. RNDr. Štefan Porubský DrSc.
Co-investigator:
Anotation:
The aim of the project is to develop new and to generalize used methods of the analytic and combinatorial number theory employed in the study of distribution and metric properties of number sequences or their families. In the area of distribution properties the aim of study will be global characteristics of the set of all distribution functions, as well as conditions for the existence of their speci c properties indicating the distribution type and their impact on the randomness and the arithmetics of the underlying sequences. The aim of the study will also be the interference between arithmetic characteristics (discrepancy, density etc.) of sequences and their sets of distribution functions and the sets of distribution functions of derived sequences (e.g. block, triangle, etc.), as well as the mutual relationships of these two sets. The aim of study will also be the interference between distribution and metric characteristics (as Hausdor dimension, Baires classi cation, etc) of number sequences and their arithmetic properties, as irrationality, or measures of irrationality.
2012-01-01 -2018-12-31Center of Excellence - Institute for Theoretical Computer Science
Number:P202/12/G061
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
Center of Excellence - Institute for Theoretical Computer Science (CE-ITI) is a research center for theoretical computer science and discrete mathematics. CE-ITI aims at becoming an international leader recognized world-wide, and also a driving force of Czech theoretical computer science and discrete mathematics. To achieve that, CE-ITI includes best professors, young researchers, and students from several top institutes in the Czech Republic, and integrates the following key activities: 1) conducting research of highest quality, addressing major challenges and open problems, and initiating new lines of research. 2) Educating a new generation of researchers and active searching for new talents. 3) Coordinating and fostering international cooperation, strengthening the standing of Czech computer science and mathematics in world-wide context.
2011-01-01 -2013-12-31Application of artificial neural networks in systems for person
Number:TA01010490
Investigator: Ing. František Hakl CSc.
Co-investigator:
Anotation:
Main goal of this grant proposal is a complex developement of localization systems (aka RTLS) in buildings covered by a WIFI network. The system will autonomously sense c hanges of radio environment. According to these changes it will tune its intrinsic parameters to keep the quality of the localization on the predefined level. This system will be used in logistic centers, production areas and in large hospitals. The application od the system will allow better usage of technological means, i.e. measuring devices, and their better utilization especially in time critical applications. e.g. in cases of emergency.
2011-01-01 -2013-12-31Interactions, information transfer and complex structures in the dynamics of changing climate
Number:P103/11/J068
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
Methods of complex networks and graph theory, tools from information theory and synchronization of nonlinear systems will be used in development of mathematical mathods and computer algorithms for analysis of high-dimensional, nonlinear time series. Spectral and Monte-Carlo Markov-Chain approaches will be used for identification of hierarchical structures and clusters in climate networks constructed from spatio-temporal fields of meteorological variables. Information flows among important clusters will be quantified using information-theoretic functionals. Structures, clusters and information flows in network representation will be confronted with known modes of climate variability and with teleconnection patterns - global communication in the Earth atmosphere. Stability of the climate system and its sensitivity to external influences (greenhouse gases, solar and geomagnetic activity) will be evaluated within the complex network paradigm. New theoretical and computational approach will help to understand natural influences and ahthopogenic forcing of global change.
2011-01-01 -2013-12-31Learning of functional relationships from high-dimensional data
Number:P202/11/1368
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is to contribute to the development of theoretical analysis of large-scale connectionistic computational systems suitable for processing high-dimensional data. The research will be focused on investigation of model complexity of systems such as neural networks, radial and kernel models in dependence on the dimension of their input data. Estimates of complexity will be derived using methods from theory of nonlinear approximation with a variable-basis and functional analysis. Based on theoretical properties, new algorithms for machine learning of connectionistic systems will be proposed as hybrid methods of constructive and population-based algorithms. The role of meta-learning for setting free parameters of the methods will be investigated. The proposed algorithms will by impelented and tested on benchmark and real data.
2011-01-01 -2015-12-31Algebraic Methods in Proof Theory
Number:P202/11/1632
Investigator:Ing. Rostislav Horčík Ph.D.
Co-investigator:
Anotation:
It is a recent trend in proof theory in non-classical logic to employ more algebraic methods and to develop the so-called algebraic proof theory. A typical example of a result in this direction is proving cut elimination in the same way that closedness w.r.t. Dedekind-MacNeille completion is proved in algebra. Other results in this direction show that current proof theory, based on Gentzen sequent calculi, works only for logics eith structural axioms of low complexity. Our project aims at generalizing the current methods and thus widening the applicability of (generalized) Gentzen calculi. A related target is a study of computational complexity of nonclassical logics.
2011-01-01 -2015-12-31Nanostructures with transition metals: Towards ab-initio material design
Number:P108/11/0853
Investigator:
Co-investigator:doc. Dr. Ing. Miroslav Rozložník
Anotation:
Development of new functional nanomaterials requires understanding nanoscale rules and mechanisms. Our aim is to elucidate the interdependence between magnetism and structure in nanostructures which contain transition metals with unfilled 3d and 4d orbitals. To achieve this we will perform ab-initio calculations of selected properties of nanostructures and observe how these properties change if the shape, size and composition of the nanostructure is varied. Our approach is based on combining calculations performed within the Green's function formalism with calculations performed by means of the finite elements method (FEM). As a by-product, this will also result in further improvemets of FEM so that it could be used in ab-initio materials science more widely. Our work will facilitate experimental and technological research aimed at practical use of nanomaterials, especially in data storage devices.
2010-01-01 -2012-12-31
Number:OC10047
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
2010-01-01 -2012-12-31
Number:ME10023
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
2010-01-01 -2012-12-31Decompositions of matrices with binary and ordinal data: theory, algorithms, and complexity
Number:P202/10/0262
Investigator:
Co-investigator: Ing. Dušan Húsek CSc.
Anotation:
The proposed project focuses on decompositions of matrices with binary and ordinal data and their utilization in dimensionality reduction of binary and ordinal data. The main goals of the project include: development of theoretical foundations for the decompositions and the calculus behind the decompositions, with emphasis on the aspects related to design of algorithms; development of efficient algorithms for the decompositions, with emphasis on algorithms based on using fix-points of closure operators and attractors of associative neural networks as factors for decompositions; development of computational complexity results regarding the algorithms for and problems involved in the decompositions; experiments and relationship to related methods of data analysis. Main outputs of the project are publications in premier journals and conferences in the field. The project focuses on basic research but has a potential to foster development in applied disciplines of science and engineering as well as a development of applications in various fields.
2010-01-01 -2012-12-31Efficient Handling Non-linear Numerical Constraints Arising in Automated Reasoning about Rich Models of Computer Systems
Number:OC10048
Investigator:doc.Dipl.Ing. Stefan Ratschan Dr.-techn.
Co-investigator:
Anotation:
The goal of the project is the design of theory, algorithms, and software for efficiently handling non-linear numerical constraints arising in automated reasoning about rich models of computer systems. The whole COST action deals with rich models of computer systems. These models are rich because they comprise a wide spectrum of logical theories for modeling purposes. One of these theories is the theory of real numbers that is especially important for the modeling of embedded systems. The project deals with this part of the whole COST action.
2010-01-01 -2012-12-31Interoperability of Future Web Services in the Web of Linked Data
Number:P202/10/P604
Investigator:RNDr. Roman Vaculín Ph.D.
Co-investigator:
Anotation:
In the project we address two critical aspects of application integration and interoperability of future light-weight Web services, namely process mediation of Web services operating in the Web of linked data, and discovery of data-centric Web services. The First problem that deals with resolving incompatibilities between Web services based processes is solved by employing techniques of AI planning and machine learning. We also address the issue of employing the user provided directions to improve the mediation process. The second problam focuses on representation and discovery of data-centric Web services. We propose a representation technique based on RDF Views over a mediated schema known from database integration area which allows to overcome the problem of data and schemas heterogeneity. In both cases, we will develop efficient and scalable algorithms with the goal of minimizing human effort necessary for achieving the dynamic interoperability.
2010-01-01 -2012-12-31Logic-based fuzzy mathematics
Number:P103/10/P234
Investigator:Mgr. Libor Běhounek Ph.D.
Co-investigator:
Anotation:
Logic-based fuzzy mathematics is a new approach to fuzzy set theory, enabled by recent advances in mathematical fuzzy logic. It features the rigorous methodology of formal logic, generalization of known results in fuzzy mathematics by automatic fuzzification of all defined notions, transferability of definitions and proofs from classical mathematics, and applicability of proof-theoretical and mode-theorecal methods. Foundations of logic-based fuzzy mathematics, laid in the past few years by members of the Prague school of mathematical fuzzy logic and comprising mainly the theory of fuzzy sets and relations, have already attracted some interest among fuzzy mathematicians (dozens of citations, awarded papers and invited talks). The aim of the project is to advance the development of logic-based fuzzy mathematics, esp. in the areas of fuzzy functions, numbers, metrics, measures, and quantifiers, and apply the apparatus of formal fuzzy logic to gradual notions of classical mathematics. The investigation, though theoretical, is relavant to most areas of applicability of fuzzy sets.
2010-01-01 -2012-12-31Model-Driven Evaluation of Design Desicion Impacts in Software Engineering
Number:P202/10/J042
Investigator:
Co-investigator:Ing. Lubomír Bulej Ph.D.
Anotation:
The development process of software intensive systems still lacks in the ability to predict the impact of early design-phase decisions on the final product. An absence of this ability - common in other engineering disciplines - keads to increased cost due to relying on a trial-and-error approach, where an incorrect design decision is often revealed only when the software is available in an executable form and any remedy is expensive. The goal of the project is extending the ability to evaluate the impact of early design-phase decisions before an executable form of a software is available - ideally when still in the design phase. The analysis and comparison of individual design alternatives will be based on the software architecture models and the analysis and simulation models, relying on both established and novel model-driven-development techniques (MDD).
2010-01-01 -2012-12-31Web Semantization
Number:P202/10/0761
Investigator:
Co-investigator:Ing. Roman Špánek Ph.D.
Anotation:
Main idea of our project is foundational research to obtain new insight into the problem of web semantization and to create an experimental environment containing - small, but sufficient portion of Web of Today - WoT (several Terra Bytes), - a portion of a future automatically Evolving Semantic Web (ESW), - several automatic semantic enrichment processes - extraction and annotation, semantic web services, ontological engineering and improving indexing infrastructure, - a reputation system designed according to the special needs of the experimental environment and a security subsystem treating possible attacks and access control in the environment WoT based on a repository, - ESW based on semantic repositories (Trisolda) and the process of semantization (an evolving chain of more and more semantic enrichments).
2010-01-01 -2014-12-31Mathematical Fuzzy Logic in Computer Science
Number:P202/10/1826
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
Mathematical fuzzy logic is a symbolic (mathematical) logic with a graded notion of truth, to be distinguished from fuzzy logic in broad sense which is a highly applied domain using the notion of fuzzy sets. Mathematical propositional and predicate logics have exact notions of formulas, axioms, proofs, semantics (standard with the real unit interval as the set of truth values and general algebraic semantics). Since Hajek's 1998 monograph, mathematical logic has been under intensive international development, in which our Czech group has played a very important role. The present project will continue this development and international cooperation, stressing the purely theoretical logical study as well as its application in computer science. In particular, the following topics will be investigated: proof theory and model theory of fuzzy logic, its relation to other logical systems, comutational and arithmetical complexity, fuzzy logic in the semantic web (description logic) and several other logics of importance for compter science (epistemic, deontic, dynamic, etc.).
2010-01-01 -2014-12-31NoSCoM: Non-Standard Computational Models and Their Applications in Complexity, Linguistics, and Learning
Number:P202/10/1333
Investigator:doc. RNDr. Jiří Šíma DrSc.
Co-investigator:
Anotation:
NoSCoM is a basic research project combining several complementary subfields of theoretical computer science in order to deepen aur knowledge of important specialized non-standard models of computation. Its aim is to characterize the computational power and efficiency of these models and to explore their ability to model problems from selected application domains such as complexity theory, linguistics, and machine learning. Our multi-model approach will also include the analysis of new resources or measures of effective computation which are introduced in the respective non-standard computational models. In particular, the research will focus mainly on non-uniform amofphous computing systems modeling mobile sensory nets, nano-scale machines, formal models of practical neurocomputing in AI, specialized unbounded automata and grammars with important applications in linguistics (e.g. word-order freedom measures), branching programs for studying space-complexity aspects of computing, models for approximating distributions of empirical data such as random forests.
2009-10-01 -2012-10-01Social Network of IT Specialists in Regions of Czech Republic
Number:CZ.1.07/2.4.00/12.0039
Investigator:
Co-investigator: RNDr. Stanislav Žák CSc.
Anotation:
2009-07-01 -2012-06-30Greenhouse gases emission reduction using information technologies
Number:M100300904
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
Processing GHG data poses a number of problems. To overcome such problems, advanced means for data processing and methods for information extraction must be developed and deployed. The missing data recovery has to be based on a formalized statistical (inevitably quite complex spatio-temporal) model. We want to develop a model for “missingness” status and check whether it is dependent on available covariates (meteorological characteristics, pollution levels, etc.). Resulting model will be (together with the GHG data) used to estimate both the missing GHG and estimate the uncertainty in the estimates. Further, we will develop and test the methods for dynamics and evolution of different variables over time, where their mutual relationships and interactions is evaluated by modern methods of analysis of time series, developed under the theory of complex nonlinear systems. Another area is mathematical modeling using numerical weather prediction (NWP) and chemical transport models operating at a sufficiently fine spatial and temporal resolution. Models of the atmosphere can be used for estimating locations of emission sources. We will test and compare the developed methods using the data from two different areas: Canada and Central Europe.
2009-03-04 -2012-02-28Preconditioning and iterative solution of saddle-point problems
Number:M100300902
Investigator:prof. Ing. Miroslav Tůma CSc.
Co-investigator:
Anotation:
Efficient solution of many tasks of computational mathematics such as construction of systems from mixed finite elements as well as problems from numerical optimization often lead via explicit or implicit solvers of saddle-point problems. The project is aimed at understanding of linear-algebraic aspects of such strategies. It also covers their applications, and solving general problems of algebraic preconditioning of iterative methods which may constitute parts of practical solvers.
2009-03-04 -2012-03-03Towards deeper understanding of Krylov subspace methods
Number:M100300901
Investigator: RNDr. Petr Tichý Ph.D.
Co-investigator:
Anotation:
Krylov subspace methods represent one of the most important classes of iterative methods for solving large and sparse linear algebraic systems. We propose to continue our work on analysis and understanding of these methods on the level of international collaboration with J. Liesen, V. Mehrmann and R. Nabben from TU-Berlin. We will focus on difficult questions and plan to investigate in particular: principles of Krylov subspace methods, mathematical theory of optimal short recurrences, convergence behaviour and related problems, model reduction and dynamical systems and solving linear systems arising from Shishkin mesh discretization.
2009-01-01 -2010-12-31
Number:ME09110
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
2009-01-01 -2010-12-31Model complexity of large fuzzy rule-based systems and neural networks
Number:MEB 040901
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The Czech partner will carry out the analysis and development of the neural networks predominantly while the Hungarian partner will do the same for fuzzy systems. During the mutual visits these results will be integrated and unified.
2009-01-01 -2011-06-30Universe of informatics
Number:CZ.2.17/3.1.00/31162
Investigator: Europe Generation
Co-investigator: Mgr. Roman Neruda CSc.
Anotation:
2009-01-01 -2011-12-31Development of Methods for Solving Large Scale Nonlinear Programming and Nonsmooth Optimization Problems
Number:201/09/1957
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:
Anotation:
The proposed project concerns development of efficient methods for solving large scale nonlinear programming and nonsmooth optimization problems. These are in particular interior point methods and nonsmooth equation methods for solving nonconvex nonlinear programming problems and also problems with equilibrium constraints (MPEC). Further, it concerns bundle methods for minimization of general nonsmooth functions and methods for solving special nonsmooth problems, among them generalized minimax problems, minimization of a sum of minima, or minimization of PC^1 functions belong. Great emphasis will be put on problems which arise in solving complicated game models, especially on problems with equilibrium constraints (MPCC, MPEC). For this purpose, special recursive quadratic programming methods and interior point methods will be developed.
2009-01-01 -2011-12-31Methods of Artificial Inteligence in GIS
Number:205/09/1079
Investigator:
Co-investigator: Ing. Dušan Húsek CSc.
Anotation:
2009-01-01 -2012-12-31Logical Foundations of Semantics
Number:401/09/H007
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
The aim of the project is to continue the previous successful project "Logical foundations of semantics and knowledge representation", i.e. to assemble a team of graduate students whose dissertation projects concern various aspects of semantics, understood as an interdisciplinary enterprise on the border of philosophy, logic, linguistics and computer science. The phenomenon of meaning is remarkable especially in that it is closely connected with what distinguishes man as a rational being. Within the framework of the project we are interested especially in the new possibilities of the approaches to meaning that have been brought about by the development of modern logic and the formal languages. These can serve as certain „models“ of natural language and consequently as means of explicating meaning. The project should especially interconnect students who concentrate on the technical aspects of this range of problems with those who are interested especially directly in the notion of meaning and who are able to critically reflect the employment of formal means for the purposes of its explication. This interconnection should effect especially an essential improvement of the dissertations of the individual members of the team, but also other results which will - thanks to the truly interdisciplinary milieu of the team - open up new ways of viewing meaning.
2009-01-01 -2012-12-31Res Informatica
Number:201/09/H057
Investigator:
Co-investigator:prof. RNDr. Jiří Wiedermann DrSc.
Anotation:
The current computer science has a so broad scope that it is hard for scientists from different sub-areas to understand each other and to communicate and exploit results of close but different sub-areas. This barrier slows down research progress due to a "reinventing wheel" effect and so it is important to re-establish awareness about results and challenges of various sub-areas of computer science especially in the early stage of research career. The proposed project "Reds Informatica" targets the above communication gap and it aims at widening horizons of PhD sudents beyond the topic of their thesis. The project involves selected students of all branches of the PhD Program "Computer Science" provided at the Faculty of Mathematics and Physics of the Charles University in Prague in collaboration with the Inst. of Computer Science of the Academy of Sciences of the Czech Republic. The project fosters collaboration between the local students via joined meetings and seminars and it provides a financial support to students to attend international meetings and to establish new personal contacts there. The joint project will also strengthen education in general scientific methodology that is critical for all students independently of their study branch.
2008-07-01 -2011-06-30Logical Models of Reasoning with Vague Information
Number:ICC/08/E018
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
One of the current tendencies in disciplines related to human reasoning is a transition from the normative approach (what is a correct reasoning) to a descriptive one (ways the reasoning actually works). This approach is influential in both applied disciplines(database theory, informatics in general) and theoretical ones (belief revision, logic). Considering the fact that in an everyday communication a majority of concepts lacks sharp boundaries, the research of vague information is undoubtebly an important part of this trend. The aim of the EUROCORES project LogICCC is to explore logical, mathematical as well as computational aspects of successful communication with structured, logically copplex, but vague information. The Czech part of the project covers the logico-mathematical part (representation of vagueness by means of formal fuzzy logic), AI part (data extraction) and logico-philosophical part (relation between fuzziness and probability, game theory as a tool of modeling vagueness). The research in this area will be performed in cooperation with the European partners of the project and will continue a long-term close collaboration of the institutions involved.
2008-07-01 -2011-12-31Úloha folátů v etiopathogenesi metabolického syndromu
Number:NS 10036-4/2008
Investigator:
Co-investigator:prof. RNDr. Jana Zvárová DrSc.
Anotation:
2008-02-01 -2009-12-31
Number:1F84B/042/520
Investigator:Ing. Mgr. David Coufal Ph.D.
Co-investigator:
Anotation:
2008-01-01 -2010-12-31Complexity of perceptron and dernel networks
Number:201/08/1744
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:prof. Ing. RNDr. Martin Holeňa CSc.
Mgr. Roman Neruda CSc.
RNDr. Petra Vidnerová Ph.D.
Anotation:
The goal of the project is to contribute to theoretical understanding which properties make networks with perceptron and kernel computational units efficient and flexible tools for learning from high-dimensional data. Estimates of model complexity of networks will be derived in terms of smoothness and oscillatory properties of data, their dimension and type of an activation function or a kernel. By inspection of these estimates, measures of data complexity with respect to various types of computational units will be proposed and characterized in terms of special norms tailored to perceptrons and kernels. Relationships of these norms to norms defined in terms of smoothness (such as Sobolev and Bessel norms) will be described. Estimates of network complexity will be derived using tools from nonlinear approximation and optimization theory. On the basis of constructive proof techniques fuzzy rules will be derived and learning algorithms will be proposed, analyzed, implemented and tested on benchmarks.
2008-01-01 -2010-12-31Decentralized control and communication
Number:IAA200750802
Investigator:
Co-investigator: Doc. Ing. Mgr. Petr Klán CSc.
Anotation:
The aim of the grant project is to derive suitable methods and algorithms for decentralized networked control systems design of selected unsolved problems arising in linear and switched interconnected dynamic systems including uncertainties and delays. We shall concentrate our attention on decentralized networked control systems design of interconnected systems, decentralized design of decentralized controllers for both similar composite systems and overlapping subsystems. The system dynamics is described by differential adn difference ODEs. Deterministic parametric uncertainties are supposed. Special attention will be paid to constructive proofs, algorithms and a real-time simulation verification on selected protopyte problems. The research will employ the results achieved under the grant of the Academy of Sciences of the Czech Republic IAA2075304 "Decentralized Control of Large Scale Systems", where no network communication issues in the feedback loop of systems were considered.
2008-01-01 -2010-12-31Intelligent Middle Agents for Mediation of Semantic Web Services
Number:ME08095
Investigator:RNDr. Roman Vaculín Ph.D.
Co-investigator:
Anotation:
The primary goal of this project is to develop algorithms for intelligent software middle agents that would automatically mediate interactions among (possibly) incompatible semantic web services. Intelligent software middle agents present a novel and elegant solution to the Web Services mediation problem. The intelligent software middle agents that can autonomously learn and adapt to the changing environments will enable on-the-fly interoperability between Web Services that cannot be achieved by existing methods without human intervention. Instead of modifying existing Web Services manually, which is very costly, the middle agents will identify relevant changes in the system and perform necessary translations automatically. The middle agents will also resolve protocols mismatches and identify missing pieces of information. This in turn will significantly decrease the interoperability costs and increase the overall flexibility of the system. Additionally, the service providers will be able to focus on their core activities instead of coping with frequent changes of the invironment. The areas which will significantly benefit from such automatic mediation of Web Services, include commercial B2B and B2C applications, governmental portals and services or distributed grid systems for scientific computations.
2008-01-01 -2010-12-31Neural Networks Learning Algorithms Based on Regularization Theory
Number:KJB100300804
Investigator:RNDr. Petra Vidnerová Ph.D.
Co-investigator:
Anotation:
The aim of this project is to study and develop algorithms for learning, namely learning approaches based on kernel methods, regularization networks and RBF networks. Regularization networks benefit from very good theoretical background, but may be not feasible for bigger tasks and their performance is very sensitive to the choice of kernel function and the regularization parameter. RBF networks, on the other hand, represent a cheaper alternative with variety of learning alhorithms, but often based on heuristics. The main goal of this project is the proposal, implementation, and experimental examination of alternative learning alhorithms. The work will be based on results of regularization theory, available optimisation algorithms and numerical algorithms. The focus will be on the weak points of the existing alhorithms, i.e. the choice of optimal network architecture, the control of generalisation ability, and the applicability on tasks with huge data sets.
2008-01-01 -2010-12-31Verification of Hybrid Systems - Exploiting the Synergy with Underlying Constraint Solving Technology
Number:201/08/J020
Investigator:doc.Dipl.Ing. Stefan Ratschan Dr.-techn.
Co-investigator:
Anotation:
The objective of the project is to design theory, algorithms, and software for the verification of hybrid systems. The motivation for studying hybrid systems comes from the fact that in modern technical systems, embedded computing devices are ubiquitous. These can be modeled using the notion of a hybrid system, that is, a dynamical system with both a continuous and a discrete state space, whose evolution is governed by differential in(equations), and discrete automata, respectively, and with an additional formalism that describes the interaction between these two components. The project will allow a break-through in the area, by exploiting the synergy between hybrid systems verification and the underlying constraint solving technology. This will be possible due to the collaboration of a Czech project group with expertise in constraint solving and numerical computational methods with the German project AVACS on the "automatic verification and analysis of complex systems".
2008-01-01 -2011-12-31BrainSync - Large Scale Interactions in Brain Networks and Their Breakdown in Brain Diseases
Number:7E08027 (7. RP EU)
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
The long-term goal of this project is to understand how neuronal assemblies exchange information (functional or neuronal communication), and how variability in neuronal communication explains variability in behavioural performance, both in the intact and injured brain. Communication involves temporal interactions between neuronal assemblies either locally within an area or large-scale between areas. We concentrate on large-scale interaction that occur at two different temporal scales: 'slow' (< 0.1 Hz) fluctuations of the blood oxygen level dependent (BOLD) signal easily measured with functional magnetic resonance imaging (fMRI), and 'fast' (1-150 Hz) neuronal oscillations measured at high (multi-unit activity (MUA), local field potential (LFP)) or low (electroencephalographz, EEG; magnetoencephalographz, MEG) spatial resolution. We wish to demonstate that these two phenomena are mechanistically linked and are behaviorally significant. A potentially important clinical application is the development of easy-to-use diagnostic measures of neuronal communication for many brain diseases such as stroke, traumatic head injury, multiple sclerosis, and Alzheimer's disease.
2008-01-01 -2011-12-31Extraterrestrial effects on atmospheric circulation in mid and high latitudes
Number:IAA300420805
Investigator:
Co-investigator: RNDr. Milan Paluš DrSc.
Anotation:
The ongoing project, in which we have shown that solar and geomagnetic activity statistically significantly affect some aspects of the Northern Hemisphere tropospheric circulation, stimulated new scientific questions, answers to which are the basis of the proposed project. Investigated will be effects of solar and geomagnetic activity on further characteristics of tropospheric circulation, i.a. cyclone positions and tracks, blockings, circulation regimes; various classifications of atmospheric circulation; bacic circulation characteristics in the Southern Hemisphere; reconstructed circulation indices over Europe back to the beginning of the 18th century. We will also study tropospheric effects of strong geomagnetic storms, energetic solar protons, and relativistic electrons, and interactions of solar and geomagneric activity with the ENSO, QBO, and stratospheric vortex. Theoretical and numerical models of downward controls of the upper atmosphere will be verified on the observed data.
2008-01-01 -2012-12-31Applications of Methods of Knowledge Engineering in Data Mining
Number:201/08/0802
Investigator:
Co-investigator:prof. Ing. RNDr. Martin Holeňa CSc.
Anotation:
Knowledge discovery in databases (KDD) is an important part of informatics with many relations to comluter science. KDD is the analysis of (often large) observational data sets to find unsuspected relationships and to summarize the data in novel ways that are both understandable and useful to the data owner. KDD is a process that covers data and business understanding, data transformations, applications of analytical procedures and presentation of results. Both the results and the effectiveness of the process strongly depend on applied knowledge. There is knowledge concerning application area, knowledge concerning analytical procedures, knowledge concerning suitable way of prosentation of results. Very important is knowledge concerning feedbacks among particular steps of the whole process. Automation of the KDD process using available knowledge is actual research problem. Goal of the project is research and development of tools for automation of the KDD processusing methods of knowledge engineering.
2008-01-01 -2012-12-31Theory of Krylov subspace methods and its relationship to other mathematical disciplines
Number:IAA100300802
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:RNDr. Iveta Hnětynková Ph.D.
doc. Dr. Ing. Miroslav Rozložník
RNDr. Petr Tichý Ph.D.
prof. Ing. Miroslav Tůma CSc.
Anotation:
We propose to continue our work in theory of Krylov subspace methods which has in recent years led to revealing of unexpected relationships and results in seemingly unrelated areas. We consider the links between different disciplines a very powerful tool and wish to exploit them further in our work. We will focus on hard questions and plan to investigate in particular: sensitivity of the Gauss-Christoffel quadrature and its relationship to the CG and Lanczos methods, extension of the core problem theory to multidimensional case, numerical stability and stopping criteria in iterative solvers, convergence behaviour in relation to various properties of the problem and open questions in the mathematical theory of optimal short term recurrences.
2007-07-01 -2011-10-31
Number:SP/1a4/107/07
Investigator:
Co-investigator: RNDr. Kryštof Eben CSc.
Anotation:
2007-03-01 -2008-12-31
Number:MEB 6-07-17 (KONTAKT)
Investigator:doc.Ing. Petr Cintula Ph.D.
Co-investigator:
Anotation:
2007-01-01 -2009-12-31Algebraic, analytic and combinatorial number theory
Number:201/07/0191
Investigator:prof. RNDr. Štefan Porubský DrSc.
Co-investigator:
Anotation:
The aim of the project is to further develop methods used in the algebraic, combinatorial and analytic number theory and thus to contribute to a better understanding of some important aspects of the arithmetical structure of the sets of integers, rational and algebraic numbers, and their underlying structures, and to extend some of these relations to arithmetical objects build on some more abstract premises as arithmetical semigroups or divisorial structure of ordered groups. Subject of study will be various density relations of number sets, questions connected with irrationality and transcendency criteria of numbers, then the theory of uniform distribution of sequences of real numbers, and properties of arithmetic structures of sets of arithmetical functions, or their generalizations in terms of abstract arithmetical semigroups. Methods will be developed which are used for characterizations of fundamental aspects of algebraic number fields and their extensions and those describing properties of their ideal structure, or those used to study abstract ideal theory in variuos algebraic systems and in the theory of divisors of ordered groups.
2007-01-01 -2009-12-31Complexity of t-norm based logics - algebraic and proof-theroetical approach
Number:KJB100300701
Investigator:Ing. Rostislav Horčík Ph.D.
Co-investigator:
Anotation:
Complexity issues belong among the most important parts of mathematical logic. Our proposal is focused on complexity of many-valued logics, especially those based on t-norms. It turns out that it is quite natural to consider a class of many-valued logics which are based on left-continuous t-norms; so-called Monoidal t-norm Logic (MTL). There are relatively many complexity results for t-norm based logics but all these results concerns only logics with continuous t-norms (e.g. Lukasiewicz, Goedel, or Hájek's Basic Fuzzy Logic). The complexity of the logics with left-continuous t-norms (i.e., logics between MTL and BL) remain still unknown. The existing methods which have been used so far for the logics with continuous t-norms rely mostly on algebraic methods. However, these methods are not directly applicable to the case of the logics with left-continuous t-norms. In order to solve these problems, we propose a certain combination of algebraic methods and proof-theoretical methods.
2007-01-01 -2009-12-31Dynamics of system aliances
Number:IAA201240701
Investigator:
Co-investigator:prof. Ing. Mirko Novák DrSc.
Anotation:
Presented proposal of Grant Projects stems from the results of the Grant No IAA2134301 "Functional Realiability Analysis of Systems Alliances" which the mostly identical team carried out in the period of 2003/2005. The goals of the proposed project reflect the major open tasks which have been met during the work on the previous grant: I. Construction and Verification of the Dynamic Multilingual Model of Systems Alliance. II. Analysis of the Translation Processes of the Alliance Languages Owing Incomplete Grammars. III. Working out of the "Phase Sensitive" Model of the Allinace Interface. IV. Verification of the Methodologies of Systems Allinace Analysis on the Background of Genetic Approaches and Artificial Neural Networks. V. Formulating the Prerequisites and Conditions of the Systems Alliance Efficiency Utilizing the Concept of Information Power. Applications could be foreseen in transportation, telecommunication, or generally in socio-technical and social complexes.
2007-01-01 -2009-12-31Solution of large, sparse and nonsymmetric linear systems with Krylov subspace methods.
Number:KJB100300703
Investigator: Dipl. Math. Erik Jurjen Duintjer Tebbens Ph.D.
Co-investigator:
Anotation:
The proposed research project addresses certain aspects of solution of large, sparse, nonsymmetric linear systems with Krylov subspace methods. First, it considers analysis of convergence behavior of Krylov subspace methods for highly non-normal model problems. Second, it investigates applicability of a new strategy to accelerate restarted GMRES. Finally, it addresses efficient preconditioning of Krylov subspace methods in the special case of solving sequences of nonsymmetric linear systems.
2007-01-01 -2010-12-31Dynamic Formal Systems
Number:IAA900090703
Investigator:
Co-investigator:doc.Ing. Petr Cintula Ph.D.
Anotation:
The aim of the proposed project is to develop a bunch of methods exploiting dynamic approaches to non-classical logics and formal systems in general, to be applied in various disciplines including computer science, analytic philosophy, and linguistics. The project follows two main branches of research: first, we explore general tools regarding: a) dynamic non-classical logics, b) game-theoretical semantics and informational independence, c) proof-theoretical characterizations, complexity aspects, and decidability. Second, we apply these tools in specific discourses: d) interrogative discourse (questions in logic and linguistics), e) deontic and cognitive discourse (applications in artificial inteligence and knowledge representation) f) inferential discourse (applications in philosophy of language and computer science).
2007-01-01 -2011-12-31Analysis of negative impacts on driver attention
Number:18/07058
Investigator:
Co-investigator:Ing. Mgr. David Coufal Ph.D.
Anotation:
2006-03-01 -2009-12-31Centre of Biomedical Informatics
Number:1M06014
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:MUDr. Hana Grünfeldová
Anotation:
The aim of the project is to develop a protocol for the preparation of an oligonucleolytic chip with the optimal set of genes for diagnostics and prognosis of cardiovascular diseases; design of a system for decision support, computer-assisted diagnostics and individualization of health care in cardiology; design of a system for decision support in stomatology and enhancement of quality in postgraduate PhD. studies in the field of biomedical informatics. The results of the project will be accessible in the form of papers in reviewed journals, presentations during prestigious international conferences and in the form of prototypes of decision support systems for cardiology and dentistry.
2006-01-01 -2008-12-31BRACCIA - Brain, Respiration and Cardiac Causalities in Anaesthesia
Number:517133 (6. RP EU)
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
Recent developments in the analysis of synchronization and directionality of couplings for nonlinear and chaotic oscillators will be applied to study the complex interactions between brain waves and cardio-respiratory oscillations during anaesthesia. Preliminary studies have shown that marked changes occur in the inter-socilator interactions during anaesthesia. These could form new bases for measurement of anaesthesia. An interdisciplinary consortium of scientist will carry out and analyse experiments on humans and rats. The work will involve (i) further development of methods for analyses of complex interactions, (ii) measurements and data collection, iii) time series analysis, (iv) mathematical modelling of interactions between coupled oscillators, and (v) neurological, physiological and clinical interpretation of mechanism of inter-oscillator couplings during anaesthesia and consciousness. Causal relationships between the oscillatory processes will be evaluated for various stages of anaesthesia and consciousness. Protocols will be prepared for a larger scale study that will be directly applicable to anaesthesiology, cardiovascular studies, neurology and physiology, but will also be useful for reviling complex interactions quite generally, e.g. in meteorology, geology, sociology and economics.
2006-01-01 -2008-12-31Development of numerical methods for solving large scale nonlinear programming problems
Number:201/06/P397
Investigator: RNDr. Ctirad Matonoha Ph.D.
Co-investigator:
Anotation:
The suggested project deals with a development of numerical metods for solving nonlinear programming problems and nonlinear approximation. It is especially about methods for large scale and structured smooth as well as nonsmooth problems, among them interior point methods and methods based on solving nonsmooth equations particularly excel. The development of algorithms for solving large scale systems of linear and nonlinear equations is connected with it. These equations are repeatedly solved in single cycles of a proposed iteration method. A special attention is paid to an implementation of new methods into the interactive software system UFO. Effective algorithms for solving general and special nonliear programming problems and their computer realization will be the main contribution of this project.
2005-01-01 -2005-12-31
Number:2004-033-2 (BARRANDE)
Investigator: Ing. Július Štuller CSc.
Co-investigator:
Anotation:
2005-01-01 -2006-12-31
Number:2005-06-060-1 (BARRANDE)
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
2005-01-01 -2006-12-31Automated model building for fuzzy logic
Number:2005-15 (KONTAKT)
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
2005-01-01 -2006-12-31Quantification of cardiovascular interactions in health and disease
Number:1/2005-06 (KONTAKT)
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
2005-01-01 -2007-12-31
Number:NR 8323-3/2005
Investigator:
Co-investigator: Prof. RNDr. Petr Hájek DrSc.
Anotation:
2005-01-01 -2007-12-31Approximation and Learning of Multivariable Functions by Neural Networks and Kernel Methods
Number:201/05/0557
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is to contribute to the interaction between classical and newly developad mathematical ideas and challenging application domains of machine learning and neurocomputing. To describe properties which make neural networks nad kernel models efficient and flexible tools for learning reasonable approximations of high-dimensional mappings, learning with generalization capablility will be investigated as a regularized optimization problem in reproducing kernel Hilbert spaces. There will be derived mathematical results on minimization of functionals defined by data, various types of their regularization, nonlinear approximation schemes, rates of approximation, integral representations, properties of operators and their eigenvalues. These results will be applied to families of multivariable functions corresponding to various neurocomputing and kernel computational models. Their consequences will be formulated in terms of complexity of neural networks and learning. Taking advantage of constructive proof techniques, fuzzy rules will be derived from forms of kernel functions which are solutions of optimization tasks. Learning algorithms computing such solutions will be designed, implemented and tested on benchmarks.
2005-01-01 -2007-12-31Formal Concept Analysis of Indeterminate and Large Data: Theory, Methods and Applications
Number:201/05/0079
Investigator:
Co-investigator: Ing. Dušan Húsek CSc.
Anotation:
The aim of the project is the development of new methods for formal concept analysis. The goal is to develop methods for analysis of indeterminate and large data and thus to contribute substantially to improvement of current methods.Formal concept analysis analysis consists in the discovery of natural concepts which are hidden in data. The data has a form of a table describing a relationship between objects and their attributes. The concepts ordered w.r.t. a natural subconcept-superconcept order form a complete lattice and form thus a hirerachical structure of clusters-concepts present in the data. The project will focus on approaches to deal with data with fuzzy attributes, to identify significant concepts, on the study of additional information constraining the concept lattice, on the study of statistical reduction of the data and the derived concept lattice, on the study of significant parts of a concept lattice, on the study of simplification/factorization of a concept lattice, and on the study of connections to other methods dealing with similar data. The project includes the development and implementation of algorithms for formal concept analysis. An important part of the project is the experimentation on large data, especially in the areas of database and information systems, software engineering, and classification of biological species. As a part of the project, a full-fledged software will be developed.
2005-01-01 -2007-12-31Formal Theories of Mathematical Structures with Vagueness
Number:B100300502
Investigator:doc.Ing. Petr Cintula Ph.D.
Co-investigator:
Anotation:
Mathematical structures containing vague concepts (fuzzy structures) have a broad range of applications; therefore they were intensively investigated during the last decades. The project is directed towards developing and applying a rich logical apparatus for the normalization of such structures. First steps in the development thereof were enabled by recent results in mathematical fuzzy logic. It is based on a simple type theory over a suitable fuzzy logic, proposed by the investigators. In the initial phase the apparatus will be completed; however, the central part of the project is its application in various fuzzy disciplines (fuzzy probability, fuzzy numbers etc.). This approach leads not only to an axiomatization, but also to a systematic study of these structures by proof-theoretic and model-theoretic methods. Moreover, the unified formalism will make an interconnection of the respective disciplines possible and will lay down the foundations of formal fuzzy mathematics.
2005-01-01 -2007-12-31New Methods and Tools for Knowledge Discovery in Databases
Number:201/05/0325
Investigator:
Co-investigator:prof. Ing. RNDr. Martin Holeňa CSc.
Anotation:
The project objective is to develop new methods and tools for KDD (Knowledge Discovery in Databases). The project is focused on the whole KDD process, from business and data understanding down to presentation of the results. That requires involvment of several branches of informatics, mathematics and integration with other knowledge technologies. This integration is understood in two ways. The first is related to utilization of the KDD results in applications of other knowledge technologies, the other is related to utilization of knowledge technologies to intensify the KDD process. The project follows up the applicant`s and co-applicant`s long-term activities. Related to these activities are internationally acclaimed results, experience with European and national grants, experience from both practical applications and teaching. The methods and tools being developed follow the line of reached results in the area of support of various phases of the KDD process. The project results shall be implemented and publicly available KDD procesures as well as appropriately published related theory. Another objective is further engagement in European research activities mainly through coordination of activities with European laboratories, which the applicants have long-term contacts with.
2005-01-01 -2007-12-31New Trends in Research and Application of Voice Technology
Number:102/05/0278
Investigator:
Co-investigator:doc. Ing. Václav Šebesta DrSc.
Anotation:
The proposed project follows up the previous research activities carried out in the speech processing area by the team that integrates all Czech research groups which are recently active in speech analysis, synthesis and recognition. It was established in 1996 to participate on an ambitious 6-year project supported by the GA CR and later continued in another speech oriented project ending in 2004. Each of the groups involved has its own proficiency in a specific domain, which allows the consortium to work on integrated and complex tasks. In the previous years the team has created large databases of annotated speech recordings, which are now available both training and testing purposes in speech recognition domain as well as for speech synthesis. In addition, a set of powerful tools and platforms for developing own recognition and synthesis systems has been built together with several working prototypes that serve for evaluation and demonstration purposes. Based on this state and with respect to the recent trends in voice technologies, the project will focus on the investigation and implementation of algorithms that are applicable in distributed, embedded and mobile systems, in recognition engines working with very large vocabularies, in TTS modules for interactive communication and information services, in automatic transcription of broadcast news as well as in multimodal audio-visual interfaces. Primarily, the research will address specific needs of Czech.
2005-01-01 -2008-12-31A new approach to knowledge representation and exchange in concept and sensor networks
Number:1P05ME799
Investigator: Doc. Ing. Mgr. Petr Klán CSc.
Co-investigator:
Anotation:
2005-01-01 -2008-12-31Collegium Informaticum
Number:201/05/H014
Investigator:
Co-investigator:prof. RNDr. Jiří Wiedermann DrSc.
Anotation:
The proposed project ("Collegium Informaticum") aims at improving and intensifying the scientific education of selected students enrolled in the PhD Program “Computer Science” at the Faculty of Mathematics and Physics of the Charles University in Prague. The research themes pertain to many areas of computer science, such as logic programming, artificial intelligence, expert systems, computational linguistics, nonstandard models of computation, software systems, combinatorial optimization and combinatorial methods and their applications. There are numerous areas of common interest among these fields, often neglected or underestimated. The project will strenghten these links stimulating new cooperation across the fields mentioned above, now and in the future. The education of doctoral students is provided in collaboration with the Institute of Computer Science of the Academy of Sciences of the Czech Republic. The grant will support participation of the students at international meetings and conferences, and it will help establish their scientific reputation and contacts. The common project and the planned joint seminars will increase interdisciplinary communication and collaboration in computer science. The project management team in which all accredited field leaders are present (together will well-known but young academicians) will assure effective project leadership and positive influence on all students of the PhD Program "Computer Science", even those who are not participating directly.
2005-01-01 -2009-12-31
Number:1M6840770004 (1M0567)
Investigator:
Co-investigator: Ing. Marcel Jiřina DrSc.
Anotation:
2005-01-01 -2009-12-31Advanced Remedial Technologies and Processes
Number:1M4674788502 (1M0554)
Investigator:
Co-investigator: Ing. Július Štuller CSc.
Anotation:
The aim of the proposed centre is research oriented on new advanced processes and technologies in reaction to the more strict requirements on underground water and rock-environment quality. The applied research will be focused on immobilization technologies and in-situ remediation, combined with the classical technologies when needed. Along with technological and processes solutions, the research will be oriented to the study of modem computer modelling methods of geochemical and biochemical processes and systems, with the purpose of improving the prediction of their time evolution and establishing available model tools for remediation management. Integral part of the Centre activities is the development of risk assessment methodologies for the inhabitans in areas subject to industrial polution or chemical accidents, reliability and risk assessment for potentially hazard technologies and risk auditions.
2005-01-01 -2009-12-31Institute for Theoretical Computer Science
Number:1M0021620808 (1M0545)
Investigator:
Co-investigator:doc. RNDr. Jiří Šíma DrSc.
Anotation:
Subjects of the research are methods, algorithms, and stuctures of theoretical computer science and their applications in information technologies, e.g. problems of large networks (WWW), problems of security and reliability, and optimalization of concrete problems. The main idea is to connect the best research centers in Czech Republic dealing with theoretical computer science and strong industrial partners for which the applied reasearch in advanced information technology is the necessary condition for further growth. From the point of view of the quality of included researchers and industrial partners it is a unique project, which is connecting the intelectual capacity of world famous scientists and economic predacity and strong capital provision of corporations, which has dominant position in the market of information technologies in the Czech Republic. Cooperation of academic and industrial sphere includes interchange of knowledge in both directions. That means application of new results in real life and also forming of new goals and priorities for basic research and training of young scientists. The center is connecting leading research communities in Prague, Brno and Pilsen, whose research interests and expertize are ideally balanced. The core of the community consists of prestigious scientists, young researchers including graduate students, who have an unrivalled conditions for a research work in the center. Organization, management and development of the center will use the experience acquired by research center ITI (project LN00A056 supported in 2000-2004). Through 5 years of its existence this center became an important center in Europe and in the world. In the personal composition there cannot be a bigger goal than to become a motiv power of theoretical computer science in Czech Republic in all of its aspects.
2005-01-01 -2009-12-31Interactive Information Portal for Algorithmic Mathematic
Number:1ET200300529
Investigator:prof. RNDr. Štefan Porubský DrSc.
Co-investigator:
Anotation:
An internet portal (gate) will be established containing expositions with study material covering various parts of algorithmic mathematics intertwined with instantaneous possibility to exploit the full computational power of computer algebra system package Mathematica for computations on various levels (of auxiliary, demo, cheking etc. character) . The internet access will open the possibility to use the computational potential of pre-programmed codes to all individuals not only those having a valid Mathematica licence. The web site will be completely maintanced by Mathematica components.
2005-01-01 -2009-12-31Mathematical Foundation of Inference and Decision under Uncertainty
Number:A100300503
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
The project is planned to be a natural common continuation of a grant project dedicated to mathematical foundations of fuzzy logics and of logics of belief and of a grant project dedicated to alternative (non-probabilistic) models and tools for managing uncertainty understood as randomness (possibilistic measures, measures with non-numerical values). The research will naturally continue in both directions, but a new twist is in systematic attention paid to the creative combination and common generalization of both. In mathematical fuzzy propositional and predicate logic, formal theory of syntax and semantics of new calculi based on the basic fuzzy logic BL will be developed; in the theory of models of randomness, various new measures will be presented and confronted with those already known. The combination of both approaches will concern fuzzy logics of beliefs, possibilistic measures for fuzzy events, testing of fuzzy hypotheses as well as non-truth functional fuzzy logics.
2005-01-01 -2009-12-31Mathematical Modelling of Natural Gas Consumption for Small and Middle Clients
Number:1ET400300513
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
The main aim of the proposed project is development of mathematical statistical models for estimation of natural gas consumption for individual consumers whose consumption is not measured on a daily or monthly basis. A further task is modelling and forecasting natural gas consumption for different segments of the gas market and for different regions. The proposed mathematical models and methods will be tested using the gas consumption measurements of selected consumers. These data will be provided by gas distribution companies in the Czech Republic.
2005-01-01 -2009-12-31Methods for Intelligent Systems and Their Applications in Datamining and Natural Language
Number:1ET100300517
Investigator:doc. RNDr. Jiří Šíma DrSc.
Co-investigator:
Anotation:
The project is focused on the development of methods for intelligent systems a their applications in machine learning, datamining, natural language processing, and computer vision. The research will benefit from a multidisciplinary approach based on the development and integration of new techniques and algorithms especially from the area of nature inspired computational models such a neural networks, decision tree, comparative inductive logic programming, Bayesian nets, extensible syntactic parsers, specialized automata, and grammars. In cooperation with commercial subjects the resulting algorithms will subsequently be implemented in practically usable systems and tested on benchmarks and real data. The project's achievements will find their use e.g. in computer linguistics, primarily in the design of grammar-checkers, natural language translators, and valency dictionaries; in computer vision (synactic pattern recognition) and information retrieval.
2005-01-01 -2009-12-31Realistic Application of Formal Methods in Component Systems
Number:1ET400300504
Investigator: Prof. Ing. František Plášil DrSc.
Co-investigator:
Anotation:
The project supports component-based application development by combining components with formal behavior description and by designing tools for automated checking of the architecture of applications composed of components with formal behavior description. The project aims to design and implement a functional prototype of a platform for formal verification of component application properties, and to propose and test methods for verification of software components and component applications using this platform. The platform will be open to the emerging methods of formal verification and code analysis, and used to test the suitability and applicability of these methods, especially with respect to model checking. The work on the formal verification methods will focus on identifying approaches to make the existing verification tools more efficient, especially in a distributed environment.
2004-10-01 -2007-12-31
Number:FT-TA/066
Investigator:
Co-investigator:prof. Ing. Ladislav Lukšan DrSc.
Anotation:
2004-09-20 -2007-09-20Complex investigation of biomechanical coditions of the artificial skeletal replacements applications, evaluatin of their failure reasons and proposal of conditions for increasing their stability in t
Number:FF-P/056
Investigator:
Co-investigator:doc. Ing. Jiří Nedoma CSc.
Anotation:
2004-07-01 -2008-12-31Application of quantum informatics on PKI (Public Key Infrastructure)
Number:1ET300100403
Investigator: RNDr. Jaroslav Hrubý CSc.
Co-investigator: RNDr. Ladislav Andrej CSc.
Anotation:
The project concerns on the application of the results in quantum informatics for the practical application in the classical cryptology and PKI from the point ov view of the information security the first of all:a) the realization of the authentization via enanglement including quantum information protocols,b) in the consequence with the PKI, digital signature and certification authorities including quantum dig. signature and the problematics of archivation from the point of view of future existence of quantum computing,c) research in quantum informatics from the new physical aspects from the point of application for communication and classical informatics,d) the research in the domain of the application of quantum game theory for cryptology
2004-07-01 -2008-12-31Information technologies for development of continuous shared health care
Number:1ET200300413
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:
Anotation:
The project will deal with the development of methods and technologies providing continuity of health care in information society. It will be based on the development of methods and techniques of remote access to information in the form of data and knowledge in connection to electronic health documentation. In the framework of the project methods and techniques supporting creation of a unified system of electronic documentation for the field of healthcare with utilization of a personal identifying chip-cards will be developed. It is very important for an applied research that its results could be veryfied in cooperation with physicians, from the healthcare experience point of view. Therefore some methods and technological solutions in the framework of the applied research will be verified in selected healthcare institutions, e.g. General University Hospital, Municipal Hospital in Caslav, Rescue Service in Caslav and selected outdoor departments of Caslav region and Prague.
2004-07-01 -2008-12-31Intelligent methods for incresing of reliability of electrical networks
Number:1ET100300414
Investigator: Prof. RNDr. Václav Snášel CSc.
Co-investigator:
Anotation:
The main objective of the research project is an increase in safety and reliability of operation of the electric power system by developing new intelligent methods. A reduction of black-out risk in the Czech Republic is very topical as well. Problems of the reliability in electric power engineering have been successfuly solved at the faculty of electrical engineering and computer science for several years. There has been established the statewide research workplace for the creation and categorisation of the databases of failures of distribution equipment and outages of supply at all voltage levels. From all over the Czech Republic and Slovakia the data are send from the particular regional distribution companies and that way is created an objectively predicative environment for the reliability computations of distribution systems.
2004-07-01 -2008-12-31Intelligent Models, Algorithms, Methods and Tools for the Semantic Web (realization)
Number:1ET100300419
Investigator: Ing. Július Štuller CSc.
Co-investigator: Prof. RNDr. Jaroslav Pokorný CSc.
Anotation:
The project at hand is a project integrating progresive theoretical disciplines of Computer Science with their practical applications. Its goal is to examine and develop theoretical backgrounds of the semantic web, mainly from the point of view of modern AI methods and methods of computing by nature, as e.i. special logics, artificial neural nets, evolutionary algorithms, datamining, etc., and to apply and verify selected methods of intelligent data and knowledge processing in that field. The results of the project are expected to be mainly:– theoretical treatises, published in international journals and presented at international conferences,– new algorithms for knowledge extraction from data, based on neural, cognitive, fuzzy and evolutionary approaches, – prototypical implementations of these algorithms within a multiagent systems. The project is centered around 12 key researchers from 3 institutions and also includes 11 Ph. D. students.
2004-07-01 -2008-12-31Mathematical modelling of air quality with applications in management of emergencies
Number:1ET400300414
Investigator: RNDr. Kryštof Eben CSc.
Co-investigator:
Anotation:
The present state of the art of information technologies enables us to simulate and forecast concentrations of atmospheric pollutants and hazardous substances in real time. Besides air quality modelling and forecasting, it is possible to model transport and dispersion of accidentally released pollutants in case of emergencies.The primary aim of the project is to apply modern informatin technologies in the field of air quality modelling and in the management of emergencies. According to the type of problem and applied methodology, the tasks are divided into two workpackages – modelling in regional scale, and modelling in microscale. A secondary task of the project is efficient dissemination of information concerning air quality to general public, and software development for the management of emergencies. The tasks of the project will be solved using modern numerical weather prediction models and chemistry transport models implemented on high performance computing platforms.
2004-07-01 -2008-12-31Modelling and simulation of complex technical problems: effective numerical algorithms and parallel implementation using new information technology
Number:1ET400300415
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:
Anotation:
The project aims at development of methods of computational mathematics and computer modelling and simulation using parallel computing and optimized computer implementations. The project will result in new (mainly iterative) algorithms of numerical linear algebra characterized by high effectivity, parallel scalability and good robustness. They will be implemented on large distributed and multiprocessor systems, and applied in particular to solution of problems of a) mechanics of underground constructions and assessment of mining technologies as well as problems of biomechanics, b) flow and transport of contaminants in porous rock including chemical reactions, c) metal forming and design of machine components, d) evaluation of realiability of complex systems, analysis of scattering of electromagnetic field of periodic structures.
2004-01-01 -2004-12-31
Number:2001-033-1 (BARRANDE)
Investigator: Ing. Július Štuller CSc.
Co-investigator:
Anotation:
2004-01-01 -2006-12-31Methods of number theory
Number:201/04/0381
Investigator:prof. RNDr. Štefan Porubský DrSc.
Co-investigator:
Anotation:
The aim of the project is to join and to coordinate the teamwork and the number theoretical research on the participating Czech reserch centers. The effort of the members of the team should concentrate to pursue further the methods of the algebraic, combinatorial and analytic number theory. More specifically, to develop new algebraic methods used for the characterization of fundamental aspects of algebraic number fields, those describing properties of the ideal structure of algebraic extensions, or those used to study abstract ideal theory in various algebraic systems and in the theory of divisors of ordered groups. The methods of the combinatorial and analytic number theory should contribute to a better understanding of the arithmetical properties (as density relations, irrationality, uniform distribution, etc.) of standard arithmetical objects as the sets of the rational and algebraic integers, or the sets of arithmetical functions, or their generalizations as abstract arithmetical semigroups. An important aspect of the activity of the members of the team should be the effort to find new fields of application of hitherto used particular methods in related or new areas and to deepen the current stage of knowledge in algebraic, combinatorial and analytic number theory.
2004-01-01 -2006-12-31New results in testing the goodness-of-fit based in Pearson-type statistics
Number:A1075403
Investigator:
Co-investigator: RNDr. Zdeněk Fabián CSc.
Anotation:
The spacings-based goodness-of-fit statistics known from the literature are shown to be asymptotically equivalent to spacings-based Pearson-type statistics. Limit laws will be extended, and the related relative asymptotic efficiencies will be evaluated for new clases of such statistics. New limit laws and the related asymptotic efficiencies will be obtained also for new clases of the Pearson-type statistics based on finite quantizations of observation spaces the sizes of which increase to infinity with the sample size tending to infinity. The asymptotic efficiences will be compared and rules for the choice of optimal test statistics will be formulated and experimentally verified.
2004-01-01 -2007-12-31
Number:A3042401
Investigator:
Co-investigator: RNDr. Milan Paluš DrSc.
Anotation:
2004-01-01 -2007-12-310
Number:0001
Investigator: Ing. Július Štuller CSc.
Co-investigator:
Anotation:
2004-01-01 -2008-12-31Development of software system for solving large-scale problems of nonlinear and nonsmooth optimization
Number:IAA1030405
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:
Anotation:
The proposed project is concerned with the development of numerical methods for nonlinear and nonsmooth optimization and their implementation in the interactive software system UFO. In particular it concerns methods for large-scale discrete and semi-infinite optimization problems containing composite partially separable nonsmooth functions and interior-point or nonsmooth-equation methods for large-scale nonlinear programming. Therefore, the development of methods for solving large-scale linear and nonlinear systems is also considered. The most complicated problems considered are problems with equilibrial constraints as nonlinear complementarity problems, variational inequalities, MPEC and EPEC. For these problems, we want to develop special methods based on interior-point and active set principles. All methods developed will be implemented in the interactive software system UFO, which will be also extended to contain new system tools, e.g., automatic differentiation.
2004-01-01 -2008-12-31Scientific-information gate for informatics and cybernetics.
Number:1N04002
Investigator: Doc. Ing. Mgr. Petr Klán CSc.
Co-investigator:
Anotation:
This project is concerned with design of a modern and open WWW portal covering needs of computer science and cybernetics in order to support research and development in these areas. Major features include effective and smart searching of scientific information and resources, auromatically updated databases, discussion forum, newsletter, articles, weblogs, authority forum and noncommercial software. The portal will stress the interaction with users. MOdern programming tools and WWW trends will be used through its development. It will be designed to be international, clean and extensible.
2003-10-01 -2007-12-31
Number:ME 701
Investigator:
Co-investigator:doc. Ing. Václav Šebesta DrSc.
Anotation:
2003-03-15 -2004-03-15Modeling the rainfall-runoff relationships by artificial intelligence methods
Number:526/03/Z042
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
The rainfall-runoff modeling is becoming more important in the view of flooding in years 1997 and 2002. The basis of our project is the application of artificial intelligence methods (namely artificial neural networks and genetic algorithms) to rainfall-runoff data from the Ploucnice river catchment. Following our previous research we will study methods based on different neural architectures. At the same time, we will test the relevance of individual model input factors as well as the length of data history. This will be done both by means of statistical methods and genetic algorithms. The created model will be also applied on data from another catchment (river Loucna), which should test its universality. All results will be statistically evaluated and compared. One of the project outcomes is also a software library realizing the developed algorithms. The resulting model will be usable for assessment and forecast of rainfall-runoff relationships, and consequently for area flood protection.
2003-01-01 -2003-12-31
Number:VaV/740/2/02
Investigator:
Co-investigator:prof. Ing. Emil Pelikán CSc.
Anotation:
2003-01-01 -2003-12-31Learning Algorithms for Local Unit Neural Networks
Number:201/03/P163
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
The main goal of our project is the proposal, design and examination of alternative learning algorithms for a subclass of feedforward neural networks with local units. The project framework consists of several theoretical results (realized by other authors as well as during our previous project) describing approximation properties and unique parameterization properties of local unit networks. We will carefully study the proof techniques of these results, try to extend them, and make use of suitable computational and numerical methods to implement them. Computational properties of the proposed algorithms, such as approximation error, convergence speed, or parallelization feasibility will be investigated. Furthermore, the practical behavior of these new learning methods will be tested on standard benchmark problems. One of the project outcomes will be a general purpose library of C++ functions—in a software package with a user interface—enabling research and engineering community to make use of our results. The algorithms will be realized in a parallel environment on a cluster of workstations.
2003-01-01 -2004-12-31BARRANDE - E-Learning a E-Business: Auto-Adapting Web Sites via Neural Networks
Number:2003-044-2
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
2003-01-01 -2005-12-31Analysis of system aliances function reliability
Number:IAA 2124301
Investigator:
Co-investigator:prof. Ing. Mirko Novák DrSc.
Anotation:
Contemporary and future systems indentified in reality are mainly of heterogenous nature. For solving of complicated tasks such systems are aglomerated into object oriented alliances. Function stability of alliance projects in alliance function realiability, determined not only by function reliability of individual alliance members, but also by conditions of mutual interactions, efficiency of alliance operations, benefit for alliance members and by resistance of alliance and all its members to disturbing influences. For the existence of the alliace these factors are of fundamental significance. The project is oriented to the analysis of mechanisms on which alliances of heterogenous systems operate, to the evaluation of their function stability and reliability and for development of methods and tools for their improvement and optimization. Attention will be paid to alliances, which as a whole have important system atributes resp. which compose hybridnor virtual system.
2003-01-01 -2005-12-31Autonomous computational agents
Number:CZE 03/023
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
2003-01-01 -2005-12-31Consortial Approach to the Development of Experimental Models
Number:102/03/0625
Investigator: Doc. Ing. Mgr. Petr Klán CSc.
Co-investigator:
Anotation:
Future progress of automatic control theory is closely linked with use of experimental models on which are possible to verify results of being theoretically developed. Getting of an experimental model is either long-term or expensive thing. Applicants therefore propose to establish a joint scientific-research consortium. Each member of this consortium will design, develop and make an appropriate physical or virtual experimental model of a real process. Another task is to derive an appropriate mathematical model and methods of automatic control based on present control theory and make this model for the other members of consortium. Project contributes to development of new experimental models based on the modern digital technologies and to obtaining of new knowledge related to automatic control theory and practice. Another contribution lies in the deeper cooperation of proposing universities and institutes. Applicants see main contribution of this consortial approach in substantial savings of both time and money on getting of more experimental models at the same time.This consortium is open for another participants.
2003-01-01 -2005-12-31Krylov subspace methods - mathematical theory, stopping criteria and behaviour in finite precision arithmetic
Number:B1030306
Investigator: RNDr. Petr Tichý Ph.D.
Co-investigator:
Anotation:
The project deals with solving systems equations (one of the basic problems of numerical linear algebra). Such systems arise e.g. from mathematical modeling of problems in sciences and engineering and they are often very large. In order to find an approximation of the solution we use iterative methods (e.g. Krylov subspace methods). To apply these methods in practice, we need to understand (among the others) principles they are based on (convergence in dependence on input data) and behaviour in finite precision arithmetic. Very important and practical questions are how to evaluate the accuracy of the computed approximate solution and when to stop the computation. We will investigate these questions. The nature of this project will require use of mathematical tools from many different areas e.g. functional analysis, perturbation theory, numerical analysis, matrix theory and numerical linear algebra.
2003-01-01 -2005-12-31Special classes of matrices
Number:A1030302
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
Investigation of special classes of matrices of the following kind: 1. Matrices determined by inequalities, such as nonnegative matrices, M-matrices and their inverses, totally positive matrices, distance matrices, etc. 2. Matrices determined by a system of equalities, or determined by a system of parameters, such as Hankel, Toeplitz, Loewner, tridiagonal, acyclic matrices, etc. Investigation of various characteristics and algebraic properties of such classes, e.g. spectral properties, as well as of mutual relationships between special classes.
2002-10-01 -2003-09-30
Number:I017
Investigator:doc. Mgr. Zdeněk Valenta M.Sc. Ph.D.
Co-investigator:
Anotation:
2002-01-01 -2003-12-31
Number:2002-030-1 (BARRANDE)
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
2002-01-01 -2004-12-31Air Quality Data Assimilation in Regional Chemistry Transport Models
Number:205/02/1488
Investigator: RNDr. Kryštof Eben CSc.
Co-investigator:
Anotation:
The main aim of the proposed grant is development of methodology of data assimilation for chemistry transport modelling, with the emphasis on regional scale and the needs of regional regulatory bodies. The CTM CAMx will be coupled with NWP model MM5 for full description of air pollution transport and chemistry of pollutants in case of smog episodes. Data assimilation techniques will be studied to improve reliability of the air pollution model results. For this purpose, the reduced rank filtering and Monte Carlo filtering methods will be investigated. The methods will be applied on selected regions of the Czech republic. Among the results a software kit will arise, which could be easily turned into a functioning system for off-line simulations of a regional CTM model. The sensitivity of the model will be studied as well with respect to the proper setting of model parameters for conditions in regions of Czech Republic, the final results will be compared to air quality data monitored in the network of the Czech Hydrometeorological Institute. For selected experiments, a comparison will be made with the results of previously used models as well.
2002-01-01 -2004-12-31Mathematical Theory of Iterative Processes with Applications
Number:201/02/0595
Investigator:
Co-investigator:prof. Ing. Zdeněk Strakoš DrSc.
Anotation:
The main goal of the project it to build mathematical theory of iterative processes for solving systems of equations arising from mathematical modeling of problems in sciences and engineering. In many application areas, understanding of iterative methods is highly desirable, but frequently not fully achieved. The computational level is often split from the modeling and discretization levels, and it is considered separately. Our project will emphasize interactions between the levels of the solution process and proceed in two complementary directions. 1.We will work on theory of iterative methods with giving emphasis to interactions between modeling, discretization and numerical computations. We will analyze mathematical properties of the methods and build up a mathematical theory of their behaviour in finite precision arithmetic. 2.We will follow the same ideas in developing, analyzing and using numerical methods in several application areas such as Markov chain modeling of reliable transportation systems, analysis of large scale dynamical systems and digital imaging. All parts of this project are tightly connected via the emphasis to sensitivity analysis of the problems to small perturbations, mathematical theory of convergence of iterative methods and thorough investigation of their numerical stability.The nature of the project will require use of mathematical tools from many different areas such as functional analysis, applied mathematics, approximation theory, perturbation theory, numerical analysis, matrix theory, graphs theory and numerical linear algebra.
2002-01-01 -2004-12-31Nonlinear Approximation with Variable Basis and Neural Networks
Number:201/02/0428
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is to contribute to the development of a new branch of nonlinear approximation theory, corresponding to approximation of multivariable functions by neural networks, and to achieve constructive theoretical results that can be applied to the concrete design of neural networks.Approximation by neural networks will be studied in a more general framework of variable-basis approximation (including also other nonlinear approximating families like free-nodes splines). Suitability of such approximation methods for solving high-dimensional tasks will be explored in terms of various norms defining sets of multivariable functions that in variable-basis approximation do not exhibit slow rates ("curse of dimensionality'').
2002-01-01 -2004-12-31Specialized Computational Models in Contemporary Computer Science
Number:201/02/1456
Investigator:prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
The project at hand is a project in theoretical computer science. Its goal is to characterize the computational power and efficiency of specialized computational models and to investigate their ability to model problems from selected application domains of contemporary computer science. The research will focus mainly to the following models and application domains: non-uniform computational models to be used in interactive, global or evolutionary computing; cognitive and neural computing models in AI; specialized finite automata and/or grammars to be used in linguistics; branching programs for studying space-complexity aspects of computing; finite computational models for approximation of empirical functions.
2002-01-01 -2004-12-31Study of Dynamical Models of Pain During the Development
Number:305/02/1487
Investigator: RNDr. Ladislav Andrej CSc.
Co-investigator:
Anotation:
The study of abnormality measures of neuronal behavior during the development in different age classes of laboratory rats for different models of pain. Unit neuronal activities of firing single neurons will be recorded in vivo at two regions of thalamus, namely in nucleus ventroposteromedialis (VPM) and nucleus ventroposterolateralis (VPL) of lateral thalamus and in medial thalamus - centrum medianum (CM), nc. centralislateralis (CL) and nc. parafascicularis (pF). Then interspike intervals (ISI) time mapping will be recovered from the measured data. The ISIs will be analyzed by methods of chaodynamics. These allow a better classification of differences between the abnormal behavior of single neurons dynamics in the case of laboratory rats with the pain and the normal ones, taken as control, in a comparison to standard methods. Besides, as it is known the neuronal dynamics is chaotic in such cases, new techniques developed for a control of chaos in general, will be applied here as a possible way of reduction of painful processes.
2002-01-01 -2004-12-31Voice Technologies a Support of Information Society
Number:102/02/0124
Investigator:
Co-investigator:doc. Ing. Václav Šebesta DrSc.
Anotation:
2002-01-01 -2006-12-31
Number:A8002202
Investigator:
Co-investigator: Mgr. Roman Neruda CSc.
Anotation:
2001-04-01 -2005-12-31COST Retional Structures in Data Mining nad Discovery Science
Number:OC 274.001
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
Further development of the GUHA method of automatic generation of hypotheses in the context of data mining. The hypotheses express general logical and statistical relations among investigated attributes and their combinations.
2001-01-01 -2003-12-31
Number:NM6548-3/2001
Investigator:
Co-investigator:prof. RNDr. Jana Zvárová DrSc.
Anotation:
2001-01-01 -2003-12-31
Number:111400003/206033
Investigator:
Co-investigator:doc. RNDr. Ladislav Pecen CSc.
Anotation:
2001-01-01 -2003-12-31Algebraic, analytic and combinatorial methods of number theory
Number:201/01/0471
Investigator:prof. RNDr. Štefan Porubský DrSc.
Co-investigator:
Anotation:
Project unifies the research capacity of the existing Czech research centers pursuing research in number theory on an international level. The aim of the project is to develop that algebraic, analytic and combinatorial methods of number theory, which are currently intensively investigated on foreign and home institutions. The methods ment are those which further develop abstract ideal theory in algebraic systems, theory of divisors in ordered groups, properties of ideal structures in various algebraic extensions and describing fundamental characteristics of algebraic number fields. Then the analytic and combinatorial methods suitable for study of arithmetical aspects as density, irrationality, uniform distribution of number sets and of properties of arithmetic functions and their generalizations on further algebraic objects like arithmetical semigroups. The aim is also to develop methods suitable for use in areas related to number theory with the emphasis on the search of new ideas for solving of known open problems. What concerns the methodological aspects, the problems will be treated using combined methods of standard algebra, category theory, mathematical analysis or combinatorics in order to deepen the knowledge in algberaic, analytic and combinatorial number theory.
2001-01-01 -2003-12-31General Asymptotic Theory of M-estimators
Number:A1075101
Investigator:
Co-investigator: RNDr. Zdeněk Fabián CSc.
Anotation:
Derivation of conditions for the consistency and asymptotic normality of M-estimators whose rho-functions are finite linear combinations of convex functions, which are practically verifiable in general statistical models with independent observations, including the linear and nonlinear regression, generalized linear models and proportional hazards models. Extension of these conditions to important types of observations on random fields. Application of the results in testing statistical hypotheses.
2001-01-01 -2003-12-31Linear Optimization Problems with Inexact Data
Number:201/01/0343
Investigator:
Co-investigator:prof. RNDr. Miroslav Fiedler DrSc.
Anotation:
Investigation of influence of inexact input data on the quality of the corresponding solutions of optimization problems is an actual research topic. In the frame of the proposed project the research activity will be devoted to three of different approaches. In the first one, coefficients of problems vary between given lower and upper bounds, i. e. the problems are formulated in terms of interval analysis. In the second approach, the data may attain values in given sets, e. g. compact, or convex ones. The third area deals with situations, in which the coeffiients are uncertain in the sense of fuzzy sets with the membership functions values from [0,1]. Here, one of the basic research tools for the investigation is the matrix theory. Special attention will be devoted to mutual relations between matrix theory, graph theory and game theory. In the literature a sufficient attention has not been given to linear optimization problems with inexact data. That is why existing results will be summarized in a unified framework and suplemented by new results. A more general form of multi-dimensional intervals, i. e. convex and compact sets, as the sets of variation of optimization data, will be investigated. Moreover, the project will deal with the inverse problems with variable data and particular algorithms for solving the above problems. As the main output of the project the monograph on the state of art will be written. The project is a continuation of the following two successful projects: "Linear problems with inexact data", (1995-97), and "Linear algebra and its applications in optimization problems with inexact data", (1998-2000), both supported by GA ČR.
2001-01-01 -2003-12-31Research of Neural Networks Capability to Provide Nonlinear Boolean Factor Analysis
Number:201/01/1192
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
The aim of the project is the development of the neural network approach to the Boolean factorization, which is assumed to be the main function of the hippocampal-neocortical interaction. Marr suggested that this function is performed by the CA3 field of hippocampus which acts as the natural neural network autoassociator. Having in mind that the brain demonstrates the higher ability to information processing comparing with artificial systems one can expect that neural network approach for Boolean factorization must dominate in practical applications when the task complexity is comparable with one performed by the brain. We suggest to use Hopfeild-like neural network autoassociator for Boolean classification of sparsely encoded factors. We will analyze the ability of this network to perform this task both by methods of statistical neurodynamics as well as in combination with computer simulation. Results of this analysis will be used in two fields. First, to model the brain ability to reduce informational redundancy of the input signals by clustering and factorization. Second, to develop the procedure for clustering and factorization of large sets of binary data. Mentioned technique can be also used in data mining area for investigation of documents similarity on Internet, for example.
2001-01-01 -2003-12-31Scalable Sparse Linear Algebraic Solvers: Analysis, Development, Implementation and Application
Number:A1030103
Investigator:prof. Ing. Miroslav Tůma CSc.
Co-investigator:
Anotation:
2001-01-01 -2004-12-31
Number:K1019101
Investigator:
Co-investigator: prof. RNDr. Milan Mareš DrSc.
prof. RNDr. Jiří Wiedermann DrSc.
Anotation:
2001-01-01 -2005-12-31
Number:ME 478
Investigator:
Co-investigator:
Anotation:
2000-03-05 -2010-05-02
Number:ghdf
Investigator:
Co-investigator:
Anotation:
2000-01-01 -2001-12-31
Number:VaV/740/2/00
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
2000-01-01 -2002-03-31
Number:IST-99-11764
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
2000-01-01 -2002-12-31
Number:201/00/1031
Investigator: Prof. RNDr. Jaroslav Pokorný CSc.
Co-investigator: Ing. Dušan Húsek CSc.
Anotation:
2000-01-01 -2002-12-31
Number:201/00/0080
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:
Anotation:
2000-01-01 -2002-12-31
Number:205/00/1055
Investigator:
Co-investigator: RNDr. Milan Paluš DrSc.
Anotation:
2000-01-01 -2002-12-31
Number:201/00/1482
Investigator: RNDr. Petr Savický CSc.
Co-investigator:
Anotation:
2000-01-01 -2002-12-312000 ADBIS-DASFAA Symposium on Advances in Databases and Information Systems
Number:HPCFCT - 2000 - 00185
Investigator: Ing. Július Štuller CSc.
Co-investigator:
Anotation:
The East-European Conference on Advances in Databases and Information Systems is the successor of the annual international workshops with the same name that during 1993 - 1996 were organized in Russia. Initiated in St. Petersburg, Russia in 1997, continued in Poznan, Poland in 1998 and Maribor, Slovenia in 1999. It became the premier database and information systems conference in the Eastern Europe, intended to increase interaction and collaboration between researchers from the East and the West, and to provide an internationally recognized tribune for the presentation of research. On the other hand, the International Conference on Database Systems for Advanced Applications was first held in Seoul, Korea in 1989 to promote database research and development activities in Asian and Australian countries. Since that time the conference has been held in every two years; Tokyo in 1991, Daejon in 1993, Singapore in 1995, Melbourne in 1997, and Taiwan in 1999. It became one of the most prestigious international conferences ever held in Asia and Australia. The 2000 ADBIS-DASFAA Symposium intended to promote interaction and collaboration between database and information systems researchers from Eastern Europe and Asia and Australia, and to provide an international forum for the presentation of research on database theory, development of advanced DBMS technologies, and their advanced applications. At the Symposium 3 invited lectures were given on the following themes: Spatio-Temporal Databases - Modelling Space and Time in Databases, An Advanced Processing Environment for Managing the Continuous and Semistructured Features of Multimedia Content and Database Systems as Middleware – Events, Notifications, Messages, besides the two tutorials on: Advanced Database Indexing and Object-Oriented Languages and Views.
2000-01-01 -2002-12-31Alternative learning procedures for feedforward neural networks
Number:B1030006
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
The majority of currently used learning algorithms for feedforward networks is based on gradient descent methods of non-linear optimization. The have proven succesfull in solving real-world problems, neverthless the are known to suffer from several shortcomings (numerical instability, time demands, communication overhead). Several theoretical results concerning the approximation power of neural networks have been establish in the last decade, but a little effort has been made to use these results - based on functional approximation theory - for proposal of novel learning procedures. We plan to carefully study the results and proof techniques of Kolmogorov, Sprecher, Kůrková, Leshno, Mhaskar and others to derive alternative learning procedures for feedforward networks. Numerical and computational properties of these algorithms will be studies by means of theory and experiments on standartd benchmarks.
2000-01-01 -2002-12-31Analysis of reliability of large hybrid technical and biological systems
Number:S1124002
Investigator:
Co-investigator:
Anotation:
Reliability and lifetime (RL) of large hybrid (i.e. containing human operators and/or users, technical artifacts or infrastructures) systems have important economical, social and environmental consequences. In order to dynamically analyze and predict RL of such systems, the methodology established on the background of system theory utilizing results of informatics, theory of system tolerances, theory of information power, as well as reliability models and standards will be carried out. The efficiency of the developed methods will be verified on samples of selected traffic and telematic company information and control system. The developed model tools will be based on progressive information technologies and concepts, identity and information power e.g. A special regard will be given to failures appearing in operation of hybrid system interfaces, where transmitted information is multilingually interpreted between the different alphabets and grammars.
2000-01-01 -2002-12-31NEUROINFORMATICS: Computational Theory of Neural Networks
Number:B2030007
Investigator:doc. RNDr. Jiří Šíma DrSc.
Co-investigator:
Anotation:
NEUROINFORMATICS is a project of a basic research in the computational theory of neural networks. One of its main objectives is to analyze the computational aspects of practically important neural network models. Along this line the computational and descriptive power of analog-state discrete- and continuous-time (symmetric) neural networks with common classes of activation functions will be studied. The learning and generalization capabilities of neural networks will be a further major research area. The complexity and approximability of loading problems will be investigated for feasible network architectures and new efficient (constructive) learning algorithms will be proposed which will be verified in the well-known PAC model. Also the techniques of neural representation processing will be analyzed within the general framework of knowledge extraction from data. The results obtained will be tested on benchmark problems and simple practical tasks.
2000-01-01 -2002-12-31Polynomial and structured matrices
Number:A1030003
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
Using the recently obtaind results on companions and infinite companions of matrix polynomials, we expect to extend the known results about matrices ocurring in linear control systems to block matrices. Applications to solving systems of linear difference equations and linear differential equations with constant coefficients are also expected. In various classes of structured matrices (e.g. Hankel, Loewner, Cauchy as well as P-matrices, M-matrices, totally nonegative matrices atc.) spectral properties, inverse eqigenvalue problems, operations within the classes and relations between the classes will be studies. Some links to geometry and some applications of the results cal also be expected.
2000-01-01 -2002-12-31Prediction of epileptic seizures: Extracting information with predictive power from scalp EEG by using nonlinear dynamical methods
Number:NF/6258-3
Investigator:
Co-investigator: RNDr. Milan Paluš DrSc.
Anotation:
2000-01-01 -2002-12-31SOFT COMPUTING: Theoretical Foundations and Experiments
Number:201/00/1489
Investigator: Prof. RNDr. Petr Hájek DrSc.
prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
The SOFT COMPUTING project is a basic and partly also applied research project in computer science and mathemtics with distinguished interdisciplinary features within both sciences. Soft computing is senn as a new integrated approach to computing and reasoning that is based on a creative fusion of emerging computational, descriptional and reasoning paradigms, such as fuzzy logic, probabilistic reasoning and massively parallel computing. Except parallel or distributed computational models the latter subsumes various non-standard models of computing, such as neuromorphic computing, genetic algorithms, cognitive computing, learning algorithms, etc. Building on the top of current status quo and on the past experience and results of its participants the main objective of the project is to provide a platform for further advancement of soft computing. This calls for a further develpoment of the respective mathematical, logical and algorithmic background, identification of potential application fields and experimental realization of selected methods developed within the project, primarily in the field of data mining, time series exploration, etc.
2000-01-01 -2002-12-31Toxic oxygen products and antioxidant eye protextion. Conditions leading to the exidative eye damage, its prevention or healing
Number:304/00/1635
Investigator:
Co-investigator:
Anotation:
The aim of the project is to elucidate condition leading to the degradation of the antiexodant eye protective mechanisms followed by oxidative eye injury and/or premature aging (such as cataract formation). Using the histochemical and biochemical approach (completed with clinical observations and computer data analysis), great attention will be devoted to changes of enzymatic antioxidants. It is supposed that degradation of enzymatic antioxidants in the anterior eye segment goes parallel with the changes of corneal and lens transparency (cataract formation) and the development of the intraocular inflammation. Furthermore, enzymes-generating toxic oxygen concentration in tears and aqueous humor. Activated cells of the inflammatory infiltrate will be studied not only as the source of reactive oxygen species, however also destructive proteases. The practical aim of this project is to introduce in clinical practice methods of detection of various enzymes in the tear fluid as markers of early anterior eye segment disturbances. The antioxidant enzymes combined with inhibitors of destructive proteases will be topically applied (dropwise) with the aim to prevent or at least diminish the oxidative eye injury
2000-01-01 -2004-12-31
Number:LN00A056
Investigator:
Co-investigator: Prof. RNDr. Petr Hájek DrSc.
Anotation:
2000-01-01 -2004-12-31
Number:LN00B107
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:
Anotation:
2000-01-01 -2004-12-31Center for Applied Cybernetics
Number:LN00B096
Investigator:
Co-investigator: Ing. Marcel Jiřina DrSc.
Anotation:
2000-01-01 -2004-12-31Mathematical foundations of inference under vagueness uncertainty
Number:A1030004
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
The project builds on the results of our preceding intensive study of many-valued logic and logics of belief that has contributed significantly to the development of their systematic mathematical and logical theory and has opened new problems as well as a way to a logical analysis of techniques of fuzzy logic (in the broad sense). This will be an advanced study of infinite-valued, modal and nonmonotonic logics (including pssobilistic and Dempster-Shafer belief logics), an analysis of fundamental mathematical theories inside these logics and development of a theory of finite models of fuzzy logic in connection with computational complexity and data analysis.
1999-01-01 -2000-12-31
Number:IGA 4870-3
Investigator:doc. RNDr. Ladislav Pecen CSc.
Co-investigator:
Anotation:
1999-01-01 -2001-12-31
Number:OK 407
Investigator:doc. Ing. Jiří Nedoma CSc.
Co-investigator:
Anotation:
1999-01-01 -2001-12-31Nonlinear approximation by neural networks
Number:201/99/0092
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
1999-01-01 -2001-12-31Parameter spaces and learning complexity of neural networks
Number:201/99/P057
Investigator: Mgr. Roman Neruda CSc.
Co-investigator:
Anotation:
1999-01-01 -2001-12-31Study of behaviour of neurons in normal states and in the pain by methods of chaodynamics - experimental and theoretical approach
Number:305/99/0049
Investigator: RNDr. Ladislav Andrej CSc.
Co-investigator:
Anotation:
1999-01-01 -2002-12-31SOFA - Software Appliances
Number:A2030902
Investigator: Prof. Ing. František Plášil DrSc.
Co-investigator:
Anotation:
The aim of this project is to design and prototype an integrated supporting electronic commerce with software components over a computer network. An associated technology will allow components to be updated dynamically inside of running systems. Business and market support facilities will provide services that allow adverising, exchange of component catalogues, matching of customer and provider profiles, component trading, etc. Based on a thorough theoretical analysis, the following issues are to be addressed: (a) Elaboration of the component provisioning and updating technology, including component versioning, dynamic reconfigurations, component transmiting, and component storage/retrieval in/from databases. (b) Elaboration of component description based on interface specification, behavioral specification, versioning model, and market-specific properties specification. (c) Support for electronic commerce with software components, including adverising, profile matching, catalogues, and component trading.
1998-09-01 -2001-08-31
Number:DMS-9802919
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:
Anotation:
1998-01-01 -1998-12-31
Number:3767-3
Investigator:doc. RNDr. Ladislav Pecen CSc.
Co-investigator:
Anotation:
1998-01-01 -1998-12-31Research in application of the object paradigm in distributed systems
Number:204 59 7436
Investigator: Prof. Ing. František Plášil DrSc.
Co-investigator:
Anotation:
1998-01-01 -2000-12-31
Number:201/98/P108
Investigator:doc. Dr. Ing. Miroslav Rozložník
Co-investigator:
Anotation:
1998-01-01 -2000-12-31
Number:404/98/1565
Investigator:
Co-investigator: Mgr. Roman Neruda CSc.
Anotation:
1998-01-01 -2000-12-31
Number:201/98/0222
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
1998-01-01 -2000-12-31
Number:OK 367
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:
Anotation:
1998-01-01 -2000-12-31Computational Models and Complexity of Computation
Number:201/98/0717
Investigator:prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
1998-01-01 -2001-12-31MGT - Medical Guideline Technology: representing, interpreting and sharing cost-effective standards
Number:OK 408
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:
Anotation:
1998-01-01 -2002-12-31Alternative mathematical models for uncertainty quantification and processing
Number:A1030803
Investigator: Doc. RNDr. Ivan Kramosil DrSc.
Co-investigator:
Anotation:
The goal of hte project is to follow up with the results of two similarly oriented projects realized in 1992-94 and 1995-97 and sponozred by the GA AS CR, in order to continue the investigation of mathematical models for uncertainty and vagueness quantification and processing based on theoretical foundations alternative to the classical probalistic and statical models. Thes concern, namely, the Dempster-Shafer theory, possibilistc theories, logics and measures, and measures with nonstandard, or in nonstandard way taken, real values. Also non-numerical, in particular, Boolean-valued uncertainty and vagueness measures will be investigated and the results will be confronted with the probalistic and fuzzy models and approaches. The goal of th eproject is, first of all, to discover and demonstrate theoretical or practical advantages of particular non-traditional model from the point of view of actual problems connected with the decision making under uncertainty.
1998-01-01 -2002-12-31Parallelization of computer processing
Number:A2030801
Investigator:doc. Ing. Václav Šebesta DrSc.
Co-investigator:
Anotation:
Design, analysis and verification of algorithms for parallel processing and efectivity evaluation of time-consuming computing tasks. A soecial emphasis will be placed on the utilization of parallel systems HP-Convex SPP 1200 in Prague and DEC Alpha in Pilsen, which are immediately available. The respective research will processed along the following main lines: 1} volume data visualization with practical application in medicine (computer tomography) and in the visualization of scientific and research results. 2) numerical linear algebra, especially the solution of large systems of linear equations, eigenvalues and eigenvector computing, preconditioning, etc. 3) neural network learning with parallelization of learning patterns, clusters of neurons or the combination of both of these approaches. 4) stochastical methods with application in the design of optimal permitted production deviations in mass production. A special part of the proposed project will be the design of benchmarking methodology for the evaluation of the above mentioned tasks.
1998-01-01 -2003-12-31
Number:RN 1998 2003011
Investigator: RNDr. Ladislav Andrej CSc.
Co-investigator:
Anotation:
1997-01-01 -1997-12-31
Number:E1030701
Investigator:
Co-investigator:
Anotation:
1997-01-01 -1998-11-30
Number:VaV/520/2/97
Investigator:
Co-investigator:prof. Ing. Emil Pelikán CSc.
Anotation:
1997-01-01 -1998-12-31
Number:OK-304
Investigator:prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
1997-01-01 -1998-12-31
Number:OK 304
Investigator:prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
1997-01-01 -1998-12-31Application of Neural Nets for Triggering of Events in Elementary Particle Physics, principal investigator
Number:Barrande 97052
Investigator: Ing. František Hakl CSc.
Co-investigator:
Anotation:
1997-01-01 -1999-12-31
Number:A1075709
Investigator: RNDr. Zdeněk Fabián CSc.
Co-investigator:
Anotation:
1997-01-01 -1999-12-31
Number:201/97/0885
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
1997-01-01 -1999-12-31
Number:205/97/0921
Investigator: RNDr. Milan Paluš DrSc.
Co-investigator:
Anotation:
1997-01-01 -1999-12-31
Number:IZ 4014
Investigator:
Co-investigator: RNDr. Jan Klaschka Ph.D.
Anotation:
1997-01-01 -1999-12-31Analysis of Informational and Dynamical Capatibilities of Recurrent Neural Networks
Number:Barrande 99010-1
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
1997-01-01 -1999-12-31Extensions of linear-algebraic problems
Number:A1030701
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
Problems of linear algebra will be extended in three directions: the coefficients of the problems may be matricial blocks, or belong to a noncommutative ring, or in the third direction, problems invloving positive definite quadratic forms will be generalized to positive multiquadratic forms, positive quartic forms, etc. The first direction is important for special classes of matrices such as M-matrices, or matrices occuring in linear control systems, such as Hankel, Loewner, companion matrices of polynomials etc., the second e.g. for totally positive matrices, the third could have prospective use in technical and theoretical-physycs problems in cases where positive quadratic forms do not suffice in the modelling.
1997-01-01 -1999-12-31Inconsistency resolution methods in the integration of data and knowledge bases
Number:201/97/1070
Investigator: Ing. Július Štuller CSc.
Co-investigator:
Anotation:
The analysis of known inconsistency conflicts of the integration of data or knowledge bases, and of existing ways of their resolution, ought to lead to the design of more general methods, namely from the theoretical point of view, but also applicable in real life situation. The focus will be concentrated on some important types of problems: the integration of data/knowledge bases with differnet priorities (hierarchically organized, centralized and decentrallized, federated, atc.), temporal inconstistemcy (its classification and resolution), the integration of data/knowledge bases with different kinds of fuzziness (vagueness, degree of truth) and of uncertainty (probabilistic, possibilistic, degree of belief). While in the case of the first two types, the formalism of the classical (predicate) logic aought to be sufficient, the last type of problems will surely require application of some variants of many-valued logics. The main goal of the grant is to formalize and to design real life applicable resolution methods for the inconsistency in the process of the integration of the data/knowledge bases. Direct applicability in envisaged in the field of classical database/knwoledge systems.
1997-01-01 -1999-12-31Preconditioned Iterative Methods for Linear Algebraic
Number:A2030706
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:
Anotation:
This project is devoted to the characterization of convergence of the Krylov space methods for large and sparse systems of linear algebraic equations as well as to the rounding error analysis of their implementations in a finite precision arithmetic. We will focuse also on the parallel implementation and on the analysis of preconditioning techniques based on various types of incomplete factorizations of the matrices and their inverses. We assume a close international cooperation, direct application of our results for solving real-world problems and the serious constribution of the research group to the education of both undergradueate and graduate students. The proposed project represents a straightforward continuation and further development of our research and teaching activities from the past few years.
1997-01-01 -2000-12-31
Number:OC 15.10
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
1996-01-01 -1998-01-01
Number:102/96/0183
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31
Number:201/96/0918
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31
Number:OK 158
Investigator:doc. Ing. Jiří Nedoma CSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31
Number:IGA MZ 3773-3
Investigator:
Co-investigator:doc. Ing. Václav Šebesta DrSc.
Anotation:
1996-01-01 -1998-12-31
Number:3763-3
Investigator:doc. RNDr. Ladislav Pecen CSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31
Number:IZ 3704
Investigator:
Co-investigator: RNDr. Jan Klaschka Ph.D.
Anotation:
1996-01-01 -1998-12-31Approximation of functions and neural networks learning algorithms
Number:201/96/0917
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31Complexity of Continuous Models of Neurocomputing
Number:A2030602
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
Complexity of neural networks will be characterized in mathematical terms of estimates of rates of convergence in approximation of continuous and measurable functions. For networks with various types of units (e.g., perteprons with various types of real and complex activation functions, radial-basis-function units with various types of kernels) and architecture (e.g., multilayer or cascade) rates of approximation will be studied in dependence on the number of the network connections and units and the size of the network parameters. Consequences of theoretical results will be formulated to help to develop a systematic design procedure for continuous neural network modelling.
1996-01-01 -1998-12-31Information Technologies Education and Training IT EDUCTA
Number:OK 188
Investigator:prof. RNDr. Jana Zvárová DrSc.
Co-investigator:
Anotation:
1996-01-01 -1998-12-31Mathematical foundations of inference under vagueness and uncertainty
Number:A1030601
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
Distinguishing sharply between fuzziness (vagueness, degree of truth) and uncertainty as degrees of belief, logical systems are being developed and investigated that formalize both. The main formal means are meny-valued logics and (generalized) modal logics, understood as systems of mathematical logic and reflected to formal methods of mathematical logic. The study includes various kinds of many-valued propositional and predicate logic, the latter with generalized quantifiers; the theories of uncertainty invloved include probability, possibility theory, and theory of belief functions. A unified exact mathematical theory is the main goal. Applications to inference in expert systems and to automated data analysis are expected.
1996-01-01 -1998-12-31Mathematical modelling of the transport and the reactions of the chemical substances in the contaminated underground water
Number:205/96/0921
Investigator:prof. Ing. Miroslav Tůma CSc.
Co-investigator:
Anotation:
The enormous contamination of the underground water in the cenoman chalk formation near Stráž pod Ralskem is one of the most serious ecological problems of the Czech Republic. Creation, analysis and application of mathematical models is the necessary condition for the long-term prediction and control of the contaminated area during mining, leaching and remediation processes. We propose to analyze and solve the porous media flow and diffusion problems simultaneously coupled with the reactions of chemical species and also to solve the inverse problem for identification of filtration coefficients. Based on the results of the two-stage solution of the chemical transport solved in our previous grant project GA ČR 201/93/0067 we will investigate the mathematical models satisfying the specific geologic and hydrogeologic conditions. Consequently, we will concentrate on the mixed and mixed-hybrid formulations. We will investigate solution of the resulting discretized nonsymmetric systems with symmetric and nonsymmetric blocks. The project assumes the coordinated solution of all the problem parts with application of results for this ecological problem.
1996-01-01 -1999-12-31
Number:IZ 3703
Investigator:
Co-investigator: RNDr. Jan Klaschka Ph.D.
Anotation:
1996-01-01 -2004-12-31International cooperation in ATLAS Detector in CERN
Number:RP-4210/69
Investigator: Ing. Marcel Jiřina DrSc.
Co-investigator:
Anotation:
1995-01-01 -1996-12-31
Number:201/95/0979
Investigator: Ing. František Hakl CSc.
Co-investigator:
Anotation:
1995-01-01 -1997-12-31
Number:A1030504
Investigator: Doc. RNDr. Ivan Kramosil DrSc.
Co-investigator:
Anotation:
1995-01-01 -1997-12-31
Number:201/95/1484
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
1995-01-01 -1997-12-31
Number:201/95/0992
Investigator:
Co-investigator: RNDr. Ladislav Andrej CSc.
Anotation:
1995-01-01 -1997-12-31
Number:201/95/0976
Investigator:prof. RNDr. Jiří Wiedermann DrSc.
Co-investigator:
Anotation:
1995-01-01 -1997-12-31
Number:308/95/0304
Investigator:
Co-investigator:doc. Ing. Jiří Nedoma CSc.
Anotation:
1995-01-01 -1997-12-31New methods of prediction and optimization for applications in electric power network of the Czech republic
Number:102/95/1311
Investigator:
Co-investigator:
Anotation:
Research and develpoment of new nonlinear aproaches and methods of prediction and optimalization with special regards to artifical neural networks applied for middle-time and long-time prediction ranges and for complication conditions. The interest will be given also to situations with limited and/or non-stable input data streams. The founded methods will be applied for the design of synthesis procedures for predictors and optimizers wich applicability will be verified on real data sets monitoring the electric power production and distribution needs in the Czech Republic.
1994-01-01 -1996-01-01Numerical Methods for Linear Algebraic Systems with Application to Nonlinear Problems
Number:A230401
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:prof. Ing. Ladislav Lukšan DrSc.
prof. Ing. Miroslav Tůma CSc.
Anotation:
It is proposed to analyze and develop methods, both direct and iterative, for solving linear algebraic systems that arise in many areas of scientific computing. Based on the results of our previous work, we will analyze Krylov space methods and consider possible extensions that choose approximate solution from outside the Krylov space. We attemp to explain the effects of finite precision arithmetics on these algorithms and propose new, more numerically stable and effective variants. We will use connections and describe implementations to classical areas as the theory of numerical integration, orthogonal polynomials and continued fractions. In the field of sparse direct methods, we will concentrate on data structures and treshold pivoting strategies; we will propose a new direct acyclic graph model. Based on it, we will attemp to reduce the size of the intermediate fill-in in the multi frontal approach, and a new hybrid left-looking/right-looking direct solver will be proposed. The results obtained for both direct and iterative methods will be used for designing new methods in the closely related research areas as sloving large structured nonlinear systems, nonlinear approximation and nonlinear programming as well as nonlinear regression. Potential for paralellism will be studied and parallel implementations will be design for specific machine architectures. Developed software will be used for the solution of real world problems.
1994-01-01 -1996-01-01Structured matrices
Number:A130407
Investigator:prof. RNDr. Miroslav Fiedler DrSc.
Co-investigator:
Anotation:
Principal directions of research - A) Theoretical background and properties of some classes of structured matrices (Hankel, Toeplitz, Vandermonde, Cauchy, Loewner, Bézout, companion etc.) and generalizations of these classes to the confluent cases and infinite-dimensional classes. B) Algorithmic problems and complexity considerations for structured matrices. C) Special subclasses of structured matrices. D) Appliactions of structured matrices. - The results will be presented in the form of publications in mathematical journals.
1994-01-01 -1996-12-31Analysis of Informational Capabilities of a Class of Artificial Neural Networks to Optimize their Structure
Number:201/94/0729
Investigator: Ing. Dušan Húsek CSc.
Co-investigator:
Anotation:
Investigation of informational and dynamical properties of associative memory based on artificial neural networks both by mathematical analytical attemp as well as combination with computer simulation. For development of new NN associative memory will be taken into account last neurophysiological investigation as a basis for hypothesis how associative memory is organized in the human's (monkey's) brain. Optimization structure of NN with regards of obtained results.
1994-01-01 -1996-12-31Application of modern mathematical methods in economic information processing
Number:201/94/1327
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
The main aim of this project is to investigate the applications of the most up-to-date methods of artificial intelligence in economy. The theoretical research within the project will be oriented on the artificial neural networks (ANN) and knowledge-based systems (KBS). The fruitful integration of ANN and KBS will result in applications of developed methods in various economic areas (e.g. exchange rate forecasting, credit risk assessment etc.). Project provides basis for combining research and educational activities of the team in connection with the collaboration with other institutions and universities and allows also research students participation. The side-effect of this project is also bringing new information know-how into economic analyses in our country.
1994-01-01 -1996-12-31Design of Modular Artificial Neural Networks
Number:A230404
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:doc. RNDr. Ladislav Pecen CSc.
Anotation:
The main goal of proposed project is to study the principles leading to design of neural network architectures with a high degree of modularity. Our interest is focused on application areas, where complex data collected from various information sources are available. Using special strategies and optimization techniques we expect the demonstrable advantages of the developed modular neural network architectures for application in medicine, energetics and economy, where some cooperation and common projects between research team and other institutes have been started.
1994-01-01 -1996-12-31Information Retrieval from Textual Databases Bsed on AT and NN Methodology
Number:102/94/0728
Investigator: Ing. Dušan Húsek CSc.
Co-investigator: Prof. RNDr. Jaroslav Pokorný CSc.
Anotation:
MAin goal of the presented project is to develop the theory and design of new methods for retrieval in textual databases with broader objective to use them in multimedial databases. It is supposed to utilize last knowledge in the area of text modelling. At nerual networks and last but not least in linguistics. As result should be data structores and algorithms which will enable achieve textual database which will communicate with the user by fully different way that do present ones products.
1994-01-01 -1996-12-31New Approaches to Neural Networks in Digital Signal Processing for Applications to System Identification and Modelling
Number:201/94/0130
Investigator:
Co-investigator:prof. Ing. Emil Pelikán CSc.
Anotation:
The project goal is in the interdisciplinary research of general digital signal processing methods and new architectures of neural networks, adaptive methods and discrete transform. Basic interest is in analysis and identification of nonlinear system and nonstationary discrete signals resulting from engineering, biological and economical structures. Proposed algorithmic solutions will be applied for the analysis and modelling of physiological signals including detection and classification of selected signal components for diagnostic purposes and they will be also used to predict behaviour of power systems and to analyse financial time series. Project provides basis for general methods study, their application in important priority areas and it allows research students to participate as well to combine reserch and educational interests of both groups in connection with their collaboration with other foreign institutes and universities. Final suggested methods and algorithms will be prepared also for application in other areas of medicine, energetics and economy.
1994-01-01 -1998-12-31Development of the psychological method of newborn behavior assessment
Number:IZ 1807
Investigator:
Co-investigator: RNDr. Jan Klaschka Ph.D.
Anotation:
1993-10-01 -1995-09-30Construction of methods for data analysis in epidemiological studies, detection of risk factors and disease risk modelling
Number:313/93/0616
Investigator:
Co-investigator: RNDr. Kryštof Eben CSc.
Anotation:
The main aim of the project is to contribute to development of appropriate methods for data analysis from epidemiological studies, for detection of risk factors and modelling of the disease risk. The Charls University in Prague, Academy of Science of Czech republic and several health institutions. In frame of the project data analyis of oncological and imunogenetic studies will be provided and theoretical approaches based on data analysis developed and adequate models using methods of mathematical statistics, logic and information theory constructed. The results achieved in the project will be incorporated into pregraduate and postgraduate education at Charles University and in the project EuroMISE of the TEMPUS programme. For practical applications in medicine and health care a prototype of the system of programs for personal computers will be developed.
1993-09-01 -1995-08-31
Number:102/93/0912
Investigator:
Co-investigator: Ing. Marcel Jiřina DrSc.
Anotation:
1993-09-01 -1995-12-31
Number:201/93/0781
Investigator:
Co-investigator: Doc. RNDr. Ivan Kramosil DrSc.
Anotation:
1993-07-01 -1995-06-30Hadron interaction in TeV region and their fast triggering for the quark-physics
Number:202/93/0178
Investigator:
Co-investigator:
Anotation:
The study of the pp interaction at the central of mass energy of 16 TeV within the international collaboration in ATLAS experiment. The experiment is under preparation and has already been admitted as a project of LHC-CERN. Our goals: 1.Computer simulation of interactions and of the full detector response concerning the physics of CP-violation and the b-quark and t-quark properties, 2.Tracking system and charged particle identification in the inner detector ATLAS, 3.The study and the proposal of the architecture for the fast triggering on interactions with a given topology based on neural networks method, 4.Develpoment and pilot technical implementation of the optoelectronical neural networks for the second-level trigger.
1993-07-01 -1995-12-31Approximation of functions and architectures of neural networks
Number:201/93/0427
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is to develop a mathematical background for choice of neural networks architecture (type of units and their activation functions, their number and interconnection) from the standpoint of the theory of approximations of functions. There will be used methods from functional analysis for investigating approximation capabilities of various types of architectures popular in applications and coding theory for optimal representation of output signals.
1993-07-01 -1995-12-31Research of optimization methods and develop of an interactiove system for universal functional optimization
Number:201/93/0429
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:
Anotation:
The main purpose of our research is a theoretical and practical investigation of optimization methods and development of optimization algorithms based on modern approaches such as sparse matrix technology, parallel processing, automatic differentiation and knowledge bases. Our attention is mostly focused to nontraditional fields including sparse and structured nonlinear programming, nonsmooth optimization and parameter estimation of dynamical systems. The new results will be used to the development of the interactive system for universal functional optimization which becomes to be a significant tool for various applications in a technical and economical practice.
1993-01-01 -1995-01-01
Number:A230106
Investigator: RNDr. Kryštof Eben CSc.
Co-investigator:
Anotation:
1993-01-01 -1995-12-31
Number:201/93/0067
Investigator:
Co-investigator:prof. Ing. Miroslav Tůma CSc.
Anotation:
1993-01-01 -1995-12-31
Number:101/93/0430
Investigator:
Co-investigator:
Anotation:
1993-01-01 -2000-12-31The structure and dynamics of relationship between the psychological profile and embodiment of students of medicine
Number:IZ 1407
Investigator:
Co-investigator: RNDr. Jan Klaschka Ph.D.
Anotation:
1992-02-01 -1994-12-31Non-numerical uncerianty quantification and processing in computer-aided systems for conclution drawing and decision making
Number:17555
Investigator: Doc. RNDr. Ivan Kramosil DrSc.
Co-investigator:
Anotation:
Processing real data and information of various kinds and proveniences for the sakes of conclusion drawing, decision making and adequate goal/oriented activity, we have to consider uncertainty and vagueness connected with the data in question, first of all, we have to investigate, how this uncertainty projects or cumulates into the uncerianty connected with the resulting conclusion, descision or action. A mathematical theory of uncertainty processing for these sakes can be based on a numerical uncertainty quantification (as is the case of probability theory, e.g.), but for a number of reasons we may consider also a non-numerical uncertainty quantificatiob and processing. The goal of the project consists in a systematical investigation of possibilities and advantages of non-numerical uncertainty quantification in conclusion drawing and decision making systems dealing with uncertainty as well as in designing such particular models for these sakes, which prove some theoretical or practical advantages, if compared with the numerical ones, and which will be practically applicable in an important degree.
1992-01-01 -1993-01-01
Number:24453
Investigator:
Co-investigator: RNDr. Ladislav Andrej CSc.
Anotation:
1992-01-01 -1994-12-31The approximation capabilities of multilayer neural networks
Number:23057
Investigator: RNDr. Věra Kůrková DrSc.
Co-investigator:
Anotation:
The goal of the project is to investigate the functional epproximation capabilities of multilayer feedforward neural networks of various types to provide a theoretical background for a methodology of determining neural network architecture. We propose to take advantage of tools of approximation theory and fucntional analysis to investigate, what properties of the function being implemented play a role in determining the type of architecture, the type of activation or kernel function and the number of layers, and to identify rates of convergence of and approximation in dependence on the number of hidden units and dimension of the input space. Furthermore we are going to study the possibilities of the simplification of learning by means of partially "hard-wired" architecture.
1991-04-01 -1993-12-31Model search techniques
Number:23003
Investigator:doc. Ing. Václav Šebesta DrSc.
Co-investigator:
Anotation:
The goal is to improve substantially the effectivenes of data-driven model search techniques by suing parallel programing and choosing an appropriate starting set of models. This is conditioned by a complexity analysis including possible heuristic techniques. For the real efectivity in applications the complexity of results is substantial as well; this problem can be attacked via appropriate methods of represenation and by evolving new evaluation criteria. The relationship between knowledge representation techniques for knowledger systems and data-based design techniques will be taken into account.
1991-03-01 -1992-12-31Mathematical foundations of inference in expert systems
Number:130108
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
It is commonly recognized fact thet the inference in expert systems is not of strictly logical and includes uncerainty processing. This gives to this inference both a logical and a probabilistic character and leeds to various new alternative theories of uncertainty. The approach specofic for the suggested project can be called both tolerant and critical. The aim is a matematically deep analysis of logical, algebraic and probabilistic apsects of at least two formalismus: the classical compositional inference mechanism and Dempster-Shafer theory of evidence. We want to avoid oversimplications and biasedness typical fro most similar analyses and contribut to theoretical foundations of inference under uncertainty in expert systems.
1991-03-01 -1992-12-31Neural Nets with Neurons of Limited Number of Synapses
Number:23005
Investigator: Ing. Marcel Jiřina DrSc.
Co-investigator:
Anotation:
The main task of the project is to develop a method for the design of neural net models with a larger number of layers and a relatively limited number of synapses for each neuron. Further, to analyze theoretical as well as the practival possibilities of the realization. Next, to state the behavior of the nets of this kind, especially from the point of view of adaptive dynamics. Finally, to state the rules for the equivalence or similarity of different kinds of neural nets from the point of view of limited interconnection and give the theory of composition of large nets from partial nets.
1991-03-01 -1992-12-31The Analysis of the Generalization Abilities of Layered Neural Networks Used for Signal Processing
Number:23010
Investigator:prof. Ing. Emil Pelikán CSc.
Co-investigator:
Anotation:
The exploration of the Artificial Neural Networks (ANN) was mostly investigated by the experimental manner up-to-now. The lack of theoretical background resulting into a very practical knowledge of the ANN behavior and generalization abilities is caused by the parallelism and ninlinearities of the ANN systems. Our research is proposed to investigate the theoretical background of the ANN with respect to the theory of application of the layered ANN in the field of signal processing. The main aim is to investigate the generalization, which plays the main role in the application of the ANN, and to verify the possibilitiesof the ANN exploration (theory of application). The methods proposed are partially neural and partially classical. The very important field of the ANN exploration is the physiological and technological signal analysis and control. This domain is used like the input data for our study the results being more general.
1991-03-01 -1993-12-31Convergence and stability of conjugate gradient type methods for solving linear systems and computing eigenvalues in finite precision arithmetic
Number:23007
Investigator:prof. Ing. Zdeněk Strakoš DrSc.
Co-investigator:
Anotation:
The goal of the project is to gain a deeper understanding of loss of orthogonality and its effect on the practical rate of convergence of the Conjugate Gradient and Lanczos-type processes. This could lead to develpomend of more accurate and more robust CG-type methods, having grate impact on the solution of real problems in a wide area of applications. We start with several open questions concerning especially the clustering of Ritz values in the finite precision Lanczos alghoritm, convergence of Ritz values and its relation to the acceleration of the CG process, and the effect of initial guess on the convergence rate of these algorithms. Our methodology combines theoretical work with numerical experiemtns based on Greenbaum's results.
1991-03-01 -1993-12-31The analysis and applications of new neural networks architectures
Number:23011
Investigator:
Co-investigator:
Anotation:
The effective application of neural networks in artifical intelligence and ither areas meets often obstacles connected with the proposal of appropriate model configuration and low understanding of their abilities (like existence and speed of adaptation process) as well as specific knowledge representation. The main goal of the project is design of coupling several network architectures with fast adaptive strategy and more transparent generalization properties. Besides we would like to elaborate a mathematically based theory of their synthesis and to develop software environment for its verification. This will be oriented partly to theoretically difficult cases, partly to the realization knwlege based neural system (including comparson with similar classical expert systems). Main ideas are based on dynamical control of treshold functions, multistage synthesis of nets and on geometrical interpetation of underlying mappings. The project team has already obtained internationally presented results in this field. Based on them, the cooperation in two abroad projects has been offered.
1991-02-01 -1994-12-31Research of optimization methods and developmentof an interactive system for univerzal functional optimization
Number:23012
Investigator:prof. Ing. Ladislav Lukšan DrSc.
Co-investigator:prof. Ing. Miroslav Tůma CSc.
Anotation:
The main purpose of our research is the investigation and the development of efficient algorithms for numerical optimization based on modern programming tools such as sparse matrix technology and parallelization. Great attention is devoted to the nondifferentiable optimization, nonlinear regression, large-scale nonlinear programing, global and stochastic optimization. The new resluts obrained will be used for the develpoment of our interactive universal function optimization (UFO) system.
1991-01-01 -1992-12-31
Number:11924
Investigator: Prof. RNDr. Petr Hájek DrSc.
Co-investigator:
Anotation:
Please follow this link to the new website of the Institute of Computer Science ...