ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 1. Vol. 23. 2017

To the contents

A. A. Razorenov, Ph. D., Postgraduate, e-mail: arazorenov@hse.ru, V. A. Fomichov, Professor, e-mail: vfomichov@hse.ru, National Research University Higher School of Economics (HSE), Moscow, Russia

A New Approach to Formalization of Instructions' Semantic Processing Based on the Theory of K-representations

The paper proposes a new approach to the development of algorithms for semantic parsing of natural language instructions, using the formal means for representing the input, intermediate, and output data. The principal theoretical result is the development of an original, broadly applicable algorithm of instructions' semantic parsing. For building a semantic representations (SR) of an instruction, the algorithm uses the theory of K-representations (knowledge representations) — the only theory today being convenient for constructing SR of arbitrarily complex instructions. The algorithm also uses a compact formal model of a linguistic database (LDB) and the notion of a graph of semantic-syntactic ties (GSST). The model of LDB and the notion of GSST are modified in comparison with their descriptions in the previous article of the authors in the journal "Informational Technologies" (Moscow, 2016, no. 12). In comparison with the known algorithms, the new algorithm can handle much more complex instructions formed by several verbs with dependent words. On the other hand, the new algorithm is very compact due to the use of the results of graph theory. The proposed algorithm has been used in the design of practically useful file manager NLC-2 with a natural language interface (Natural Language Commander — Version 2) which was developed by means of Haskell programming language.
Keywords: natural language processing, semantic parsing of instructions, semantic representation, theory of K-representations, SK-language, dictionary of lexical frames, linguistic database, graph of semantic-syntactic ties, natural language control of file system, Natural Language Commander

P. 3–14

V. V. Naumova¹, Professor, Head of Laboratory, e-mail: naumova@fegi.ru, S. E. Dyakov², Scientific Researcher, e-mail: sergdkv@gmail.com, K. A. Platonov¹, Junior Researcher, e-mail: platonov@fegi.ru
¹Far East Geological Institute FEBRAS, Vladivostok
²Institute Automation and Control Processes FEBRAS, Vladivostok

Technologies of Search and Submission of Information in the Scientific Infrastructure "Geology of the Russian Far East"

The block scheme based on distinction in types of geological data: spatial, quantitative, bibliographic and expert knowledge, is an approach basis for realization of Infrastructure. In each separate information block of Infrastructure various approaches and technological decisions are applied to integration, storage and search of the data. The concept of the organization of a uniform point of access to this Infrastructure is in providing to the user all possible information on geological objects of the Far East of Russia and services of its processing at a time, in one place, in a uniform point of access. Technologies of search and submission of information for this Infrastructure are described in article.
Keywords: information infrastructure, integration of non-uniform resources of information, geological information, geology of the Russian Far East

P. 15–20

G. G. Bulychev, DSc., Professor, Moscow Technological University, Moscow

Method of Spatial Characteristics in Problems of a Mechanics of a Deformable Solid Body. Part 1

In this article the method of spatial characteristics is considered and analyzed and its possibilities for studying of wave processes in deformable solid bodies of various structure and a rheology are investigated. Merits and demerits of a method are considered, and also the circle of tasks which it is necessary to solve for building of the characteristic form of mathematical model of dynamics and dynamic destruction of researched bodies and for its usage in numerical simulation. The initial and characteristic form of dynamical equations of the anisotropic symmetrically elastic bodies is shown. By means of the considered method numerically on PC the research problem of destruction of the two-storey house is solved at an earthquake load.
Keywords: dynamic processes, a method of spatial characteristics, numerical simulation, dynamics and destruction of buildings and constructions

P. 21 – 27

M. A. Styugin, PhD, Researcher, e-mail: styugin@gmail.com, Siberian State Aerospace University (SibSAU), Krasnoyarsk

Indistinguishable Executable Code Generation Method

The classical objective of obfuscation considers indistinguishability of the obtained code in relation to original programs of equal functionality. The present paper reviews another objective of obfuscation, when indistinguishability is considered relatively to programs of different functionality. In this case, an obfuscator is provided with a key at the input. It is nearly impossible to discover the program's functionality without having the key, therefore that obfuscator type is named a key obfuscator. The paper proves existence of a key obfuscator and existence of efficient algorithms for its operation demonstrated by recursive functions. The practical relevance of the result obtained by the research is the possibility to store an executable code in an untrusted computational environment and to obstruct injection attacks and distributed computing analysis.
Keywords: obfuscation, indistinguishability, untrusted computations, key-based obfuscation, computational complexity, polynomial time algorithm

P. 28–33

K. A. Shcheglov, Graduate Student, A. Yu. Shcheglov, Professor, e-mail: info@npp-itb.spb.ru,
St. Petersburg University of ITMO, Russia

Operational Safety. Potential Offender Attack Threats Implementation Modeling

We researched attack threat exploitation parameters and characteristics elaborating modeling problem. Its parameters differ from reliability ones by taking into account potential violator interest and willingness for executing real attack threat and also taking into account attack goals. We suggest an approach for building mathematical model of potential violator, introducing quantitative complexity characteristics (measures) for attack threat implementation and for informational system danger level (from potential violator). We suggest an approach of quantitative estimation of attack threat actuality (to make decision on security implementation while designing informational security system). We research attack threat modeling problems and features while different potential violator goals (like information confidentiality, integrity and availability violation).
Keywords: attack threat, vulnerability threat, mathematical modeling, potential violator model, attack threat implementation complexity, danger from potential violator, attack threat model, attack goal

P. 34– 41

V. N. Kostin, PhD., Assoc. Prof., e-mail: vladimirkostin57@mail.ru Federal State Educational Government-financed Institution of Higher Professional Education "Orenburg State University"

Synthesis of Optimal Placement of Technical Means of the Systems of Physical Protection Critical Facilities

The method of optimum placement of the technical means of protection (TMP) at design of the systems of physical protection (SPP) of crucial objects is developed. Basic data is the plan of object ofprotection with characteristics of an arrangement of buildings (elements) and the security indicators demanded to them. Feature of a task is the different required level of criterion of security of elements of object of protection which SPP has to provide.
On the basis of the graph of penetration of the violator, using operation of composition of a matrix of contiguity of the graph, all possible ways of penetration to object elements were formed. The received results were formed in the form of an incidence matrix: lines — ways of penetration, columns — the graph's edges. The field of a matrix — Boolean variables: one if the edge of the count is included in a movement route, zero — otherwise. It is necessary to provide counteraction to penetration of the violator on all routes of penetration. By means of a task about a covering on a matrix of incidence options of suppression of penetration of the violator (danger development) were defined. Each covering is characterized by two parameters: extent and removal from the protected object element. In the first parameter determine the cost of a covering, and by the second — probability of timely arrival of group of reaction which defines time of reaction to penetration of the violator.
The method of the analysis of hierarchies has determined the acceptable TMP set for formation of SPP. On the basis of a method of dynamic programming for various levels of potentially dangerous buildings (requirements to security size) the plan of optimum placement of TMP meeting the set criteria of their security was defined. Method advantage — the important principle of design of SPP — the principle of uniform strength is provided.
Keywords: the technical means of protection (TMP), system of physical protection, an incidence matrix, a task about a covering, dynamic programming, the plan of optimum placement of TMP, criterion of security of object

P. 41–49

A. V. Skorokhodov, Researcher, e-mail: vazime@yandex.ru, V. E. Zuev Institute of Atmospheric Optics of the Siberian Branch of the Russian Academy of Sciences

Application of Viola — Jones Method for Atmospheric Gravity Waves Detection above the Water Surface on Modis Satellite Imagery

The purpose of this work is to develop an algorithm for automatic identification of cloud views of atmospheric gravity waves above the water surface on MODIS satellite imagery with a spatial resolution of 1000 m based on the Viola — Jones method. The coastal regions of the oceans with the highest repeatability of the researched events in the period from 2007 to 2015 (the coast of the Arabian Peninsula, the Madagascar Strait, the western coast of Australia, the Kurile Islands) are determined. Different views of atmospheric gravity waves on MODIS satellite imagery are shown. Paper describes Viola — Jones method, its advantages and disadvantages. Training and background samples for learning of the research events identification algorithm are formed. Proposed modified Haar features for detecting of quasi-parallel lanes of atmospheric waves views on satellite imagery. The initial size of "sliding" window at this work is 40 x 40 pixels to reduce algorithm false positives at the satellite imagery fragments of stratocumulus un-dulatus clouds. The results of atmospheric gravity waves identification above water surface on MODIS and Himawari-8 satellite imagery with a spatial resolution of 1000 m are presented. The estimate of the correct classification probability of the test samples (more than 300fragments of MODIS imagery) equals 0,87. The proposed algorithm can be used in operational monitoring of atmospheric waves because it allows to efficiently identifying cloud views of the research events with different structure on the background of the optically thick cloudiness using Nvidia CUDA technologies. The results of atmospheric waves identification on satellite imagery are discussed. The perspective development directions of this work are presented, which include the development of automatic methods to search for their generation sources, the estimation of their impact on the environment and the prediction of their occurrence over a certain territory.
Keywords: atmospheric gravity waves, classification, cloudiness, Haar's features, satellite data

P. 50–56

Yu. V. Polishuk, PhD, Associate Professor of Computer Security mathematical software and information systems, e-mail: Youra_Polishuk@bk.ru, Orenburg State University

On the Method of Monitoring Information Entropy the Large Technical Systems

The paper considers a method of monitoring the information entropy of large technical systems. Management decisions for large technical systems realizes a group of decision-makers, based on the analysis of factual data obtained from the accompanying operational content.
The need for the participation of a group of decision-makers, decision-making process is dictated by the scale of the control system and the complexity of the control process. The complexity of large technical systems, as well as the complexity of obtaining and processing of factual data accompanying operational content, contribute to the divergence of the real state of the control system and the status reflected in an operational content. The latter contributes to the growth of information entropy of large technical systems and the adoption of incorrect management decisions. Thus, the adoption of correct management decisions for large technical systems is allowed only if they meet the conditions of the maximum allowable value of information entropy.
The method of monitoring the information entropy of large technical systems on the basis of the accompanying operational content increases the efficiency of management decision-making by the automated evaluation of the completeness of knowledge about the system and the use of semistructured operational model content for collector-ray acquisition system allows you to automate this process.
Keywords: entropy of the system, semistructured data, electronic documents processing, large technical systems

P. 56–61

M. Yu. Chinkov, Student-Bachelor, e-mail: michael.chinkov@gmail.com Penza State Technological University (Penza, Russia)

Tracking Services State in Web-Application Infrastructure

This article regards the web-application services track and suggests a service discovery software as a one specific way of problem solution.
Keywords: web-application, service discovery, Consul, healthchecks

P. 62–64

A. A. Kharlamov, D.Sci., Senior Research Fellow, kharlamov@analyst.ru, Institute of Higher Nervous Activity and Neurophysiology of Russian Academy of Science, Moscow

Associative Memory as a Neuronetwork Environment of Combined Knowledge Space

An associative human memory is an environment for combined knowledge space development. Text information processing as an example of any modality information processing by human is shown. The results of the process are a homogenous network of key text concepts which are weighed by their rank in text and hierarchical thematic structure tree which characterize the text complexity. The results are interpreted clearly by certain text analysis using thematic modelling approach. The n-gram text model is used in­stead of traditional unigram (bag of words) one.
Keywords: complexity, corticoform associative memory, structured information processing, associative (homogenous semantic) network, thematic hierarchy tree, multimodal information, thematic text analysis, n-gram text model

P. 66–75

V. V. Pekunov, D.Sc, Engineer, OAO "Informatika", pekunov@mail.ru

The Derivation of the Information from Artificial Feed-Forward Neural Networks in the Form of the Simple Algebraic Models

In this work the building of the simplified analytical models of the artificial neural networks with some diminishing of accuracy is investigated. The activation functions of the neurons are replaced by polynomial expressions which are accurate at the work areas of the neurons. The importance of the different intervals of work areas is taken into account. It is proposed to use the strategy of non-increasing the maximal power of input variables in the result to obtain the more compact solution. The result combining is performed by means of the developed symbolic math kernel. This kernel uses accelerated algorithms including the parallel computations. An appropriate acceleration and effectiveness of the parallelizing are achieved on the 16-core machine of the Google's Compute Engine platform.
Keywords: artificial neural network, analytical model, simplifying, symbolic math, parallel computations

P. 76–80

To the contents