Proceedings of the Second International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’17)

The fundamentals of synergetic artificial intelligence and its relationships with swarm intelligence are considered. Basic classifications of agents and multi-agent systems are presented, the comparison between intelligent and reactive agents is made. Different synergy sources for conventional group intelligence and swarm intelligence are elicited. The concepts of swarms, swarm intelligence and swarm robotics are discussed. The principles and models of swarm tasks distribution via local interactions are formulated. The results of experimental investigation of pack-hunting task are analyzed.

In this paper, we consider the ideas and approaches to the intelligent data analysis methods in problems of information security. Solving network security problems is a complex task, involving the large number of factors and requiring finding reasonable compromises between maintaining security, the stable work, enhancing operating expenses and functional restrictions of complex information systems. There are considered the ways to apply inductive concept formation methods for analyzing network traffic, as well as argumentation methods for an automated support of security solutions. The proposed approach allows to give numerical assessments of the quality of recommendations developed by the system, thereby helping to solve an important task - the task of choosing the way to react to suspicious activity in the system. In addition, the example of handling the dangerous situations arising in the system is given.

The game approaches are rather popular in many applications, where a collective of automata is used. In the present paper the game involves a group of learning finite automata. The game is played sequentially with one automaton at a time, the result of the game defines next automaton to be played with. The goal of this game to provide some measuring system that is a reminiscent of collecting statistics in Probability Theory but in a different manner.For measuring of an unknown membership value a new concept has been introduced called Cognitive Generator which transforms a fuzzy singleton to ordinary crisp logic value. Considerations on various types of axiomatic approaches show that the Cognitive Generator, as well as our Evidence Combination Axiomatic, belongs to one class of axiomatic theories, which may be used in applications directly.Some programming examples aimed to illustrate our general approach.

An approach to the sensitivity analysis of local a posterior inference equations in algebraic Bayesian networks is proposed in the paper. Performed a sensitivity analysis of first a posterior inference task for stochastic and deterministic evidences propagated into the knowledge pattern with scalar estimates. For each of the considered cases the necessary metrics are chosen and transformations are carried out, that result into a linear programming problem. In addition, for each type of evidence theorems that postulate upper sensitivity estimates are formulated and proofs are provided. Theoretical results are implemented in CSharp using the module of probabilistic-logical inference software complex. A series of computational experiments is conducted. The results of experiments are visualized using tables and charts. The proposed visualization demonstrates the high sensitivity of the considered models, that confirms the correctness of their use.

Precise daily weather forecasts are necessary for the utilization of renewable energy sources and their penetration into grid systems. Standard meteorological statistical post-processing methods relate local observations with numerical predictions to eliminate systematic forecast errors. Neural networks, trained with the last historical series, can model the current weather frame to refine a target forecast for specific local conditions and reduce random prediction errors. Their daily correction models can process numerical prediction model outcomes of the same data types (instead of the unknown data) to recalculate 24-hour wind speed forecast series. Global numerical weather models succeed generally in forecasting upper air patterns but are too crude to account for local variations in surface weather. Long-term complex forecast systems, which simulate the dynamics of the complete atmosphere in several layers, cannot exactly detail local conditions near the ground, determined by the terrain relief, structure, landscape character, pattern and other factors. Extended polynomial networks can decompose and solve general linear partial differential equations, being able to model properly unknown dynamic systems. In all the network nodes are produced series of relative polynomial derivative terms, which convergent sum combinations can directly define and substitute for the general differential equation to model an uncertain system target function. The proposed local forecast correction procedure using adaptive derivative regression model can improve numerical daily wind speed forecasts in the majority of days.

We present the overall design of a data acquisition system developed for the needs of the SLADE (Stress Level and Emotional State Assessment Database) database. The database consists of synchronized EEG, ECG, skin temperature (ST), and galvanic skin response (GSR) recordings, used in stress level assessment and recognition of emotional states. SLADE will facilitate the development of automated tools and services for stress-level assessment and monitoring.

This paper investigates the use of SVM algorithm with PCA method in classification, which is one of the most common task of machine learning. Classification is the problem of identifying to which of a set of categories a new observation belongs, on the basis of a training set of data containing observations (or instances) whose category membership is known. SVM algorithm can produce accurate and robust classification results on a sound theoretical basis, even when input data are non-monotone and non-linearly separable. So they can help to evaluate more relevant information in a convenient way. PCA method reduces the dimensionality and the maximum number of new variables that can be obtained is equal to the original, with new variables are not correlated with each other. Experimental studies have shown that it is possible to improve the accuracy of SVM classification algorithm using PCA method.

The paper considers integrated tools based on multi-agent temporal differences reinforcement learning and statistical modules. Implemented algorithms of reinforcement learning methods are described. The possibilities of including anytime algorithms into the forecasting subsystem type of intelligent decision support system of real-time for improving performance and reducing response and execution time were proposed. The work is supported by RFBR and BRFBR.

The article describes practical application of the original method of training artificial neural network based on the poorly formalized expert knowledge. The method allows extending the range of problems to be solved in case of lack of a sufficient number of the observations due to the fact that the training vectors are formed on the basis of the expert knowledge. The expert continuously defines classes of objects that are generated by the pseudorandom number of the training vectors of input signals, and created visual images by computer for clearly describing objects by given training vectors. The method is applied to solve important practical problem for determining of the atmospheric surface layer stability. The problem is formulated as a classification problem. As being the artificial neural network was selected multilayer perceptron. This trained neural network is represented by programming model implementing as DLL-module of dynamic-link library. The research bases on the original computer program that implements the algorithm of author’s training method. The program determines and implements the steps of the author’s research using heuristic training method of the artificial neural network to solve the problems of classification on the basis of poorly formalized experts’ knowledge. Its algorithms are used to generate visual (cognitive) images of possible situations to retrieve the unconscious expert knowledge. As a result, the aim of the study was achieved. The proposed method of training artificial neural network was applied successfully to solve a practical problem and showed its efficiency on an example of the classification problem. The author’s training method is protected by Russian patent for invention; the use of computer software holds a certificate of state registration.

The article analyzes the problems that arise when creating subject-oriented intelligent systems based on information from the Internet. Approaches are described that allow improving the methods of collecting information and making an in-depth analysis of documents in order to improve the quality and accuracy of subject-oriented collections of documents. In addition, approaches are described to build self-learning systems that expand knowledge based on the collected information.

These days web resources keep and process a lot of valuable information. Confidential data and private pages have to be protected due to business processes. To implement this requirement and limit the number of people having access to the restricted resources, you need to configure a proper authentication on website. Unfortunately authentication is often implemented incorrectly which leads to information leaks. Frequently websites have only password protection and use other simple methods. The article deals with different comparison of authentication methods, using both simple and advanced approaches including cryptography and biometrics. Moreover, the authors give the comparative analysis of different approach parameters. Usability, performance, security and other features of the methods are analyzed. The most convenient to use, the easiest to implement and the most secure methods are found. A conclusion about the most suitable application areas of each method on World Wide Web resource is made. Possible combinations of approaches and their further implementation tendency are also discussed. An analysis of domestic and foreign literary sources, scientific articles and publications on our topic is made for the finding verification. We perform Russian and international literature search on the scientific databases as well as on the electronic library systems. Furthermore, patent research is made for finding practical implementations. It includes patents of the business organizations related to web authentication methods. The results of this research show the topic relevance, the increasing number of patented authentication methods on web resources, as well as a fairly high potential of the new method development. This way should base on improvement and unification of existing approaches and on developing of new original algorithms. As a result, it is concluded that further improvements in the trends are in the field of hybrid systems.

Second-order cybernetic models let explain an influence of mass behavior upon macroeconomic characteristics. In particular, we consider situations related to the self-organization and synergy of interacting socio-economic systems and an impact of random factors. In such situations catastrophic intensity of offensive adaptive mass behavior may produce a negative impact on the economic stability. Nonlinear dynamics of self-organization processes complicates prediction of macroeconomic characteristics via extrapolation of trends. An amplitude-frequency analysis of oscillatory self-organization processes let obtain more relevant forecasts.

The paper presents a method for analyzing hybrid project workflows based on the author’s RV-grammar by using BPMN as an example, as well as translations of such project workflows into languages describing business processes (e.g., BPEL with the author’s RVTt-grammar). The effectiveness of author’s grammars for hybrid project workflows’ analysis and translation at large design and manufacturing enterprises is assessed.

Authors propose the analysis of design-technology workflows at large design manufacturing enterprise, the-level structure of workflows according to IBM Rational Unified Process methodology is developed. The problem to coordinate (approve) the design-technology documentation is studied, the author’s model of a Petri net modeling standard workflows under approval of design and technological documentation is developed.

The paper proposes the approach to uniformly present complicated objects and processes with a variable structure. The approach is based on the approximation of specific models, describing the object or the process by a uniformed remodeling class. The application of this approach is useful while solving optimization and control problems. Neural network models, which proved their high approximating capability, are suggested as a remodelling class. Applying the given approach is considered on the example of modelling of inertial torque transformer (ITT) workflow. This process has a cyclical pattern, where each phase of the cycle is divided into four segments, described by various systems of nonlinear differential equations with the same parameters. Moreover, the solution to the system describing the next segment depends on the solution to the system obtained by the previous segment. It noticeably makes it difficult to determine the optimum ITT parameters, as the fit function has the solution to the system of equations, describing the last, fourth segment of the cycle. The neural network model allows simplifying the solution to the given problem for each segment of the cycle. The input layer of the remodelling neural network was supplemented by taking into account the real output values from the previous moments of time. The neural network with such kind of structure demonstrated high level of accuracy.

The problem of synthesizing the control of the intellectual suspension of a vehicle is regarded as a problem of synthesizing optimal law of the Lagrangian system parametric control. There has been developed a method for synthesizing optimal law parametric control in conditions of uncontrollable external exposures, what allows to speak of the possibility to build a “smart” damping system. The obtained laws of control are technically feasible being characterized by low amount of computational effort.

The paper presents a way of an ontological support of design thinking when a designer discovers a necessity to solve the new project task in developing a system with the software. The proposed way is aimed at finding errors, correcting and preventing them, as well as extracting of useful questions and supporting the processes of understanding in a real-time process of designing. The basis of the proposed way lays in the use of a precedent-oriented approach to work with a project task, the statement of which is built during its conceptual solution. Results of designer’s interactions with the ontology are used for its enriching. The way is materialized in the instrumental environment WIQA (Working in Questions and Answers) supporting the conceptual designing.

A hard problem of acquiring and processing complex heterogeneous information for cargo train on-line monitoring is faced. It requires the consideration of modern measurement-information techniques and tools enabling visual, acoustic, laser monitoring of freight transport on railway. The main objective of the authors is to develop an integrated intelligent post of railcar monitoring and on-line diagnostics in station. In this context intelligent measurement concepts and methods are envisaged. A three-leveled hierarchical system of measurement ontologies is introduced that includes meta-ontology, upper ontologies and lower (domain, task, application) ontologies. A top-down methodology for ontological design of measurement is proposed. It is based on both collective and granular meta-ontology that supposes the representation of ontology by algebraic system and the use of various granulation bases. Visual representations of measurement domain and measurement tasks ontologies by mind maps are suggested. A classification of measurement uncertainty factors is made on the basis of GUM standard. Hybrid fuzzy-stochastic measurement granules are constructed. Here we use both an extended Zadeh’s information granulation approach and Takagi-Sugeno model of fuzzy dynamic system.

The paper deals with the conceptual foundations of functional objects calculus as a formal theory of systems. The basic concepts and definitions of the calculus of functional objects are presented, within the framework of which the functional object is considered as a system described in terms of the systemic-object approach ‘‘Unit-Function-Object’’. The authors introduce a number of basic definitions and operations on functional objects, that are: adding, deleting the fields of a streaming object within the calculus; redefining of fields; redefining the methods of unit objects; connections of unit objects according to various criteria.

In the paper, two formalisms are compared: the step theory of Active Logic with the formalism of extended logical programs, taking into account the different specifics of their application. Particular attention is paid to the relationship between the formalism of extended logical programs and the formalism introduced by the authors of the step theory of Active Logic with two kinds of negation.

A task of sequential pattern generation can be considered as a problem which is inverse to sequential pattern mining. This paper presents two novel approaches to the sequential pattern generation with noise, namely the approach based on stochastic automata and context-free grammars and the approach based on Hidden Markov model. The distinctive feature of these methods is the suitability to produce an output in the noisy and fuzzy input data. Also, we present the detailed calculation algorithms to the proposed approaches.

Multi-service computer networks (MSCN) play an important role in the modern society life. However, design of MSCN is a rather complex challenge. Development of adaptive routing algorithms, which consider the failures of nodes and communication lines because of the impact of the computer attacks, holds a specific place in MSCN design. The paper offers a new approach to adaptive routing in MSCN based on a combined use of the multi-path routing of data streams and the integral criterion, which is based on fuzzy assessment of network states. The algorithm based on this approach considers additional routing metrics, i.e. the level of information security, the technical state of network elements and the packet loss probability. The experimental assessment of the offered algorithm of fuzzy adaptive routing showed that in the conditions of high level impact of computer attacks the gain in the time of message delay is improved by 2–4 times. It testifies about higher performance of the proposed algorithm in comparison with known algorithms.

Time series transforming is considered as a preprocessing stage in various data mining techniques. To obtain relevant and accurate result in time series analysis it is needed to apply relevant time series representation by suitable transformation. In the paper at the first time five transformations of a numerical time series derived on the basis of a single linguistic variable of time series values are described systematically. The formal notion of five kinds (fuzzy matrix, fuzzy vectors, fuzzy linguistic, numerical and linguistic) of time series produced by these transformations and general scheme of their computing are represented. Applications of these five representations of a numerical time series in data mining techniques are given and discussed.

The paper considers a new adaptive approach for anomaly detection and prediction in continuous time series. Presented approach is based on immune double-plasticity network. Antigens, which characterize temporal patterns or separated time samples, and antibodies, which characterize models for adaptive anomaly detection, are the basic classifiers in the proposed technique. The general idea of the approach is based on interaction between antigens and antibodies, when new time sample is observed. This interaction leads to the model adaptation with respect to input data based on prediction error minimization. In particular, flexible parametric fuzzy models are used as antibodies, which provide both parametric and structural adaptation simultaneously with input data processing. Appropriate trend of model parameters shows forthcoming anomaly and saltatory structure changing shows anomalous presence. Computational experiments for presented approach together with conventional regression models are illustrated.

Nowadays the developers of robot and automated systems face the problem of analysis and interpretation of different signals, which reflect physiological processes in a human body. This is connected with increasing requirements to the means of effective support of interactions between users and computers. One of the ways to solve this problem is using the models of human emotions (operator), who takes part in forming or monitoring controlling actions in an automated system. The authors propose approaches to monitoring human emotional states using assessment of a limited number of characteristics of speech samples or electroencephalograph (EEG) signals. In order to analyze and interpret these signals the authors use methods of nonlinear dynamics, which allow reconstructing an attractor using a limited time series fragment. The paper describes a test procedure and the results, which show the changes of attractor properties during the influence of auditory incentive bunches with the same emotional interpretation on an operator. The article presents a transition sequence from one attractor to another as a fuzzy time series. Each time series is based on the characteristics of one attractor (point density of trajectories in its center surroundings). The estimates of attractor sizes allow defining an emotion sign. The direction of emotional reaction development is determined based on fuzzy estimates of an attractor density increment sign, when the attractors are reconstructed for two subsequent watch windows. Fuzzy estimates of density increment of three attractors, which are reconstructed for three subsequent watch windows, determine the trend of testee’s emotional state. There follows the prediction of direction of operator’s emotional state development. The paper shows the results of using the algorithm for analysis of EEG attractors when listening to (a) one musical incentive, (б) several bunches of incentive.

The presence of partial discharge pattern in medium voltage overhead lines with covered conductors may indicate insulation fault, rupture or downfall of the line. These failures can cause problems in the electrical energy distribution to customers. This paper focuses on the detecting and classification of the partial discharge patterns. The presented method transform a captured input data into fuzzy signatures. This allows us to deal with the captured data as regular text documents. The obtained fuzzy signatures are used in the classification phase using k-NN. The proposed method can correctly classify the captured data up to 76% of accuracy.

Presence of partial discharge pattern implies the fault behavior on insulation system of medium voltage overhead lines, especially with covered conductors (CC). This paper covers the examination of Approximation and Sample entropy as a signal complexity measures on partial discharge patterns of several kinds of faults. These features are calculated on multiple different adjustments such as the different applied denoising schemes and varying embedding dimensions. The final results reveal the splitting ability of these complexity measures on applied data.

The article develops an intellectual model for evaluating the running characteristics of the cut based on the fuzzy system of production rules that reflect the expert knowledge of technologists about the relations of objective parameters characterizing the behavior of the cut at the rolling section with qualitative parameters characterizing the running quality of the cut. The hybrid hierarchical model is developed for predicting the process of rolling of the cuts on the basis of the objective parameters of the automation object and the weakly formalized parameters characterizing the running quality of the cuts.

This paper presents the results of volatility forecasting for indices of the Russian stock market using existing and developed by the authors fuzzy asymmetric GARCH-models. These models consider various switching functions which are taking into account the positive and negative shocks and are built using the tools of fuzzy numbers. Furthermore, in some models there are used switching functions that consider expert macroeconomic information. It was shown that fuzzy asymmetric GARCH-models provide a more accurate prediction of volatility than similar crisp models.

In this paper a novel approach to fuzzy hierarchical clustering of short text fragments is presented. Nowadays dataset which contains a large and even huge amount of short text fragments becomes quite a common object. Different kinds of short messages, paper or news headers are examples of this kind of objects. Authors have taken another similar object which is a dataset of key process indicators of Strategic Planning System of Russian Federation.In order to reveal structure and thematic variety, fuzzy clustering approach is proposed. Fuzzy graph as a model has been chosen as the most natural view of connected set of words. Finally, hierarchy as a result of clustering obtained as desirable presentation structure of large amount of information.

This paper proposes an approach to comparative analysis of the inference methods for systems with many inputs using fuzzy truth values. Obtain criteria of efficiency and proven their execution for some fuzzy implications in logic systems. Proofs presented for any t-norms in the generalized modus ponens rule.

The problem of optimal allocation of service centers is considered in this paper. It is supposed that the information received from GIS is presented like second kind fuzzy graphs. Method of optimal location as method of finding fuzzy base set of second kind fuzzy graph is suggested. Basis of this method is building procedure of reachability matrix of second kind fuzzy graph in terms of reachability matrix of first kind fuzzy graph. This method allows solving not only problem of finding of optimal service centers location but also finding of optimal location k-centers with the greatest degree and selecting of service center numbers. The algorithm of the definition of fuzzy base set for second kind fuzzy graphs is considered. The example of finding optimum allocation centers in second kind fuzzy graph is considered too.

It is established that when the Hamilton-Ostrogradsky principle is satisfied, the solution to the extremal control problem can be obtained on the basis of the combined maximum principle up to a synthesizing function. This allows to find a lot of controls that, with the use of fuzzy logic, can serve as the basis for the synthesis of multi-mode dynamic systems. The effectiveness of the approach is demonstrated on the basis of mathematical simulation of the solution of the problem of synthesis of control of the mathematical pendulum.

We define a fuzzy logic whose sentences are Boolean combinations of propositional variables and Allen’s relations between temporal intervals. For this logic, we present the deduction method based on analytical tableaux. We also present the method of query answering over fact bases for ontologies written in this logic.

The paper deals with one of the most significant problem of computer design – a VLSI placement problem. The authors prove a relevance of new heuristic methods for the given problem. The VLSI placement problem is formulated and an optimization criterion is suggested. As an optimization method there is suggested an algorithm based on bacterial colony behavior in nature according with environment conditions. On the basis of these observations it is developed a new bioinspired algorithm for the VLSI placement problem. There are conducted series of computational experiments for comparison of the developed algorithm with known VLSI placement algorithms on the basis of IBM benchmarks. Obtained results are represented in a table. Conducted researches shown the efficiency of the suggested approach better then 8–10% on average.

The paper considers one of the most important problem in the design field – VLSI fragments placement within restricted construction area. The VLSI fragments placement problem is NP-hard and complex problem. In the work there are presented a description of the placement problem and transition from a circuit diagram to a graph model. To reduce dimension of the problem the authors suggest a hybrid approach based on two fractals aggregation and genetic search methods. As well as there is developed a genetic algorithm that allows to obtain sets quazi-optimal solutions during polynomial time. To confirm the effectiveness of the suggested approach there is shown an example of the VLSI fragments placement solution, developed software and computational experiment. Conducted tests and experiments approve promising of the suggested approach, a time complexity of developed algorithms is represented in the best case as ≈O(nlogn), in the worst case - O(n3).

In the near future, Smart-Grid technologies will have an incredible impact on the economics of power systems and on environment. This will be possible thanks to the intelligent communication and computer systems, which would allow the system to accommodate much more energy from renewables by combining different technologies for energy storage, electric vehicles and demand response. The main contribution of this paper is the development of models for the different components of the Smart-Grid, which can be easily generalized for many different studies. The modelling framework includes energy storage, renewable energy sources, electric vehicles and demand response. We use the problem of distribution grid power balancing to illustrate the application of the models for improving the economic performance of a balancing group. The problem is formulated as a mixed-integer linear program and can help energy companies and custumers to make investment decisions for smart-grid.

In the design of lighting applications, a maintenance factor (MF) is widely used when aggravation of lighting installation parameters should be taken into account. There are many publications offering guidelines for MF value selection. Most of them suppose that predefined maintenance plan already exists and recommendations for MF value selection are given based on that. In the reality that is not a common case and the designer embeds given MF into a project without adequate notion for that how the new lighting equipment will be actually exploited. The proper selection of MF requires definite arrangement of all exploitation measures affecting lighting parameters of the equipment. In the current work the MF value is derived from the optimal set of exploitation measures. The optimality is defined as a minimum of life time exploitation cost. The task for proper MF value selection is brought to an optimization problem and all considerations connected with that are discussed in details in the paper.

Such problem of modern navigation as correction of errors of inertial navigation systems is considered. Most often use Kallman’s filter to the solution of this problem. Use of irregular exact measurements allows to solve analytically a problem of creation of the adaptive discrete filter of Kallman, and also determine to precisely such parameters of the filter as coefficient of strengthening and coefficient of adaptation, providing a zero error of estimation thereby. The example confirming sharply increase of convergence of an algorithm of estimation due to reorganization of parameters of the filter is given.

We propose new technologies, principles and mechanisms for solving the planning problem, using mathematical methods, which lay down the principles of natural decision-making mechanisms. For a compact presentation of the solution of the planning task, a Polish expression is used. This allowed to create a solution space, within the framework of which a search process based on modeling the adaptive behavior of an ant colony was organized. Compared with existing algorithms, the results are improved.

This paper deals with a new method for decomposing search domain in a global optimization problem. Proposed method was designed for parallel population algorithms but also can be used as a diversification tool in sequential algorithms. New decomposition technique was compared with a traditional approach by means of numeric experiments with a use of multi-dimensional benchmark optimization functions and Mind Evolutionary Computation algorithm. Results of the experiments demonstrate the superiority of new technique over a canonical approach which resulted in a higher quality of obtained solutions.

In this paper are presented the problems for realization of direct adaptive neural sensorless control in combination with vector principle for induction motor control. Control system containing neural controllers of the speed and flux channels and neural speed estimator is proposed. These neural controllers perform a function of both speed and active stator current controllers (for the first channel), and respectively flux and excitation stator current controllers (for the second channel) compared to classical vector control. Neural speed estimator is designed as a neural model of the plant. For the controllers and estimator are used on-line trained backpropagation neural networks. Simulation research confirmed sufficient system performance at wide range input signal variation is done.

The Locally Recurrent Probabilistic Neural Network (LRPNN) consists of an input layer, three hidden layers and an output layer. The first two hidden layers are derived from the original PNN, while the third layer referred as recurrent layer is capable to model correlations within temporal sequences of observations. In the present study, we investigate the feasibility of FPGA-based implementation of the locally recurrent layer of LRPNN. An important consideration due to the specifics of this architecture is the use of modules with very high precision in the hardware design. Although expensive in terms of available resources in the FPGA chip, this is necessary, in order to compensate for the added error of quantization due to the multiple feedbacks from neurons in the neural network. The weights for the recurrent layer of the LRPNN are automatically computed from the available training data and translated into the hardware design. The experimental evaluation was carried out on the DEAP database, where two classes of emotional states were considered. The design makes use of a computed short-term energy from a 32-channel electroencephalographic (EEG) signal as an input. Results of an extensive experimental validation show that there is approximately one percent difference between the accuracy achieved with CPU-based software and FPGA-based hardware implementation of the LRPNN.

We proceed from the assumption that the basic abstraction of systemic approach should not be specific properties and concepts, but systemic properties and concepts. In recent decades, ontology is used to structure, to formalize and to unify the representation of data and knowledge for the purpose of their reuse. The content of ontological schemes’ structures, in the form of a conceptual apparatus of formalized universally-abstract constructions (concepts, relations and transformation mechanisms), are concrete empirically obtained abstractions of reality applied to the design of concrete structures of abstract objects. When solving the problem of formalizing the process of organizing a connection between concepts, the emphasis of research is shifted from empirical processes of concepts organization to formalized processes of their creation. The aim of this paper is to create a symbolic form for the structure of a cognitive procedure abstraction, and to demonstrate examples of the logic of organizing the structures of cognitive procedures abstractions into a structure of systemic concepts and their organic unity.

In current times, malefactors chances to succeed in performing a social engineering attack on company usually depends on how much personal information about employees he owns. Thus, search and analysis of public information about company’s employees from social network websites with purpose of protection company from malicious actions is important issue. This article is devoted to methods of identifying user’s online footprint in website of social network VK.com. Prototype of the tool for identifying employees public pages using binary decision trees as classifier is presented. Approach to fully automated gathering of training dataset is described.

Goal of this study is to compare five statistical algorithms used for evaluation of software packages based on Monte Carlo techniques. These methods are the following analysis: regression, correlation and Bland Altman as well as Wilcoxon rank sum and Kolmogorov-Smirnov tests. The methods were applied for the case study of validating a dedicated computer code for calculation of scattered radiation reaching the eyes of the operator during interventional procedures. The transport of x-rays and their interactions with the matter have been accomplished by the use of Monte Carlo techniques. Results were presented in the form of number of registered photons as a function of their energy and further compared to data from literature. From the five statistical methods for comparison, the most suitable for our application turns out to be the Bland-Altman method. Further effort is related to the development of a specific software application for evaluation of data generated from general purpose x-ray simulations, as well as with analysis of more data obtained from different incident x-ray spectra.

The paper presents iterative optimization of the Monte Carlo raytracing settings for photometric analysis of the optical system of industrial light-emitting diode (LED) luminaires with lenses. The research has been done about the influence of the main parameters of the raytracing settings (number of traced rays, number of ray reactions and weight of ray) on the accuracy of the quantitative and qualitative photometric analysis results. Tabular and graphical representations of the results are shown. The optimal Monte Carlo raytracing parameters for two different lens optical system types of LED luminaires are selected.

In this paper, an application of Markov Decision Processes (MDP) for modeling selected marketing process is presented. The process is converted into MDP model, where states of the MDP are determined by a configuration of state vector. Elements of the state vector represent most important attributes of the customer in the modeled process. Movement between the states is determined by actions of the customer. In constructed MDP model, individual states with assigned initial reward values then represent consequences of action chosen by the customer resulting in either incresing or reducing the revenue following moving into these states. Value iteration method is then used for computation of the expected final rewards for each state. Based on provided realistic data, customer behavior is analyzed and the best course of action is proposed. Model suitability for future predictions of desired action outcome rate is discussed as well.

The paper considers usage of fine-tuning of the deep neural network ensemble for recognition of 60 event types in the set of 60,000 images from WIDER database. The applied ensemble consists of two deep convolutional neural networks (CNN) using the GoogLeNet architecture, previously trained on other image bases: ImageNet and Places. Separately the accuracy of recognition of 10 events was analyzed: “Car Racing”, “Ceremony”, “Concert”, “Demonstration”, “Football”, “Meeting”, “Picnic”, “Swimming”, “Tennis” and “Traffic”. During the ensemble training output layer in the each of deep CNN is replaced to the layer with respectively 10 and 60 neurons and we tune only weights which connect output layer with previous one. The classification accuracy of 10 event classes from the WIDER image database averages 83.22%, for 60 event classes accuracy is 50.4%. In addition, the approach based on the automatic features formation using deep CNN provided a much better recognition quality of social events compared to the choice of features manually (LBP, LDP or HOG) and their further classification by support vector machine. The testing time of the developed ensemble provides the possibility of using the classifier in practical applications of event recognition with a processing speed up to 20 frames per second.

In this paper an artificial neural network multilayer structure with back propagation error is synthesized for identification of seven groups of pure signals and signals with six type of noises. The signals and noises simulations are based on the LabVIEW environment. Principal component analysis is used in the process of neural network training. The experiments are done for different types of activation functions and quantities of neurons entity in any of intermediate levels. As a results for noise identification an accuracy of 92.3% is obtained.

The aim is to propose basic steganalytical tool that can use multiple methods of analysis. We describe two detection methods that were implemented. These methods include improved detection capability than conventional steganalytical tools thanks to use of artificial neural network and several other innovative improvements. In our work is important to understand the behavior of the targeted steganography algorithm. Then we can use its weaknesses to increase the detection capability. We analyze prepared stegogrammes by application of several conventional algorithms such as image difference. Then we can determine where are the most suitable areas of image for embedding the message by steganography algorithm.Two of our plug-ins are focused on steganography algorithms Steghide, OutGuess2.0 and F5. These algorithms are open source and easy accessible, so the risk of their abuse is high. We use several approaches, such as calibration process and blockiness calculation to detect the presence of steganography message in suspected image. Calibration process is designed for creation of calibration image, that represents the original cover work and for comparison to suspected image. Blockiness calculation serves us as a statistical metric that react to the presence of secret message. Next we deploy the artificial neural network to improve detection capability.Second plug-in utilizes a detection method that is based on analysis of inner structures of JPEG format. This detection method uses overall quality calculation based on quantization tables and Huffman coding table. These informations are processed by neural network that is able to decide whatever the suspicious file contains embedded data and which steganography algorithm was used to create this file with tested confidence larger than 93% and for detection capability up to 99%.

The study of physiological signals and the evaluation of their features are of great importance for the automated emotion detection, as these are directly connected with the successful modelling and classification of the states of interest. In the presented work, we present an evaluation of the appropriateness of LFCC and the logarithmic energy of signals as features for automated recognition of negative emotional states in terms of recognition accuracy. In particular, three sets of features are compared – features computed after frame-level segmentation of the signal; features computed after averaging of frame level descriptors; and features extracted from an entire EEG recording. The performance of the extracted features is evaluated using C4.5 classifier for 10, 15, 20, 30, 45, and 60 filters.

Electrical impedance tomography (EIT) is an imaging system suitable for long-term monitoring. To extend current uses of EIT, improvements in the image reconstruction algorithms are essential. New image reconstruction methods for EIT can be tested on an impedance model of human body. Moreover, accurate anatomical impedance distribution models of human body are used to generate training data used in machine learning algorithms.Simulation framework, introduced in this paper, is capable of autonomous conversion of Computed tomography (CT) scans from DICOM format into 2D MESH human thorax impedance distribution model. Developed impedance models of large thorax structures achieve accurate results through segmentation of CT images and Fourier Fitting. Framework is developed in MATLAB as an extension to EIDORS and NETGEN frameworks.

The paper describes the method of musical compositions generation by the image color spectrum based on user ratings. It describes the algorithm of image analysis, based on conversion of image from RGB space to HSV and the algorithm for generating melodies based on a correlation table between artistic and musical characteristics by J. Caivano, and the table of correlation of colors and notes by Newton. It also describes the algorithm for harmony generation based on the assessment of euphonious of chord progressions by users.

This paper investigates the problem of geoinformation modeling of situations and solutions adopted in the implementation of logistics projects. The task is solved within the framework of the geoinformation model, which displays the experience of implementation of similar projects. The performance criterion of the model is the reliability of the generated decisions. The decision is reliable if it corresponds to the reality displayed by the electronic map. The displaying method previously completed projects into a given space-time area is analyzed. The method is based on the image representation of geoinformation models. The geoinformation model of the logistics project includes a scheme consisting of logistics centers and transportation routes. The procedure of displaying of project image is described that is implemented by an intelligent component with knowledge about displaying. Principles of the reliability evaluation of the results of the transformation are discussed.

In this paper a novel method of pipes detection in images on an industrial enterprise is proposed. Detailed description of preprocessing procedures is given. Method is supposed to work with grayscale images. First, contrast of the source image is increased. Then smoothing procedure is applied to reduce noise influence. Estimation of filtering quality is performed with PSNR and MSE methods. Different edge detection operators were tested, Canny operator shows the best results. Hough transform and its fuzzy form are considered. Fuzzy Hough transform is a modification of the original method that uses fuzzy features based on fuzzy sets theory. Extracted lines are filtered or fused to decreased features number. Algorithm parameters are shown in the table with their optimal values for current task. Proposed method is implemented in C++ with the use of OpenCV library. The results of this algorithm, number of errors and executing time on test and real image are described.