ISSN 0236-235X (P)
ISSN 2311-2735 (E)

Journal influence

Higher Attestation Commission (VAK) - К1 quartile
Russian Science Citation Index (RSCI)

Bookmark

Next issue

2
Publication date:
16 June 2024

Articles of journal № 4 at 2014 year.

Order result by:
Public date | Title | Authors

11. Double-layer vector perceptron for binary pattern recognition [№4 за 2014 год]
Authors: Kryzhanovsky V.M., Malsagov M.Yu., Zhelavskaya I.S.
Visitors: 13981
A new model of neural network, Double-Layer Vector Perceptron (DLVP), to solve nearest neighbor search problem, is proposed. The problem of single-layer perceptron, when error of the even one output neuron leads to fail of all network, is solved. DLVP is improved single-layer perceptron with additional layer, which accumulates information. As result, it is possible to right recognize even if all middle layer neurons are fail, i.e. neural networks with ‘weak’ neurons can be built. The model was compared with a single-layer vector perceptron. This comparison showed that though its operation requires slightly more computations (by 5 %) and more effective computer memory, double-layer vector perceptron excels at a much lower error rate (four orders of magnitude lower). We obtained the estimate of DLVP storage capacity and analyzed its properties. During this analysis we found out that the model has one more useful property, which single -layer vector perceptron does not have, i.e. using the proposed model we can effectively solve k nearest neighbors search problem.

12. The technique and framework for dynamic vulnerabilities detection in binary executables [№4 за 2014 год]
Author: Shudrak M.O.
Visitors: 7678
The article introduces an original technique for binary executables security analysis that allows analyzing ne t-work applications security using fuzz testing which is effective for software bugs detection. Vulnerable network applications are a tremendous threat to network communications security. Vulnerabilities in web browsers and server software can cause a disastrous effect. A good example of it is HeartBleed critical vulnerability in OpenSSL TLS implementation that allows an unauthenticated remote attacker to retrieve critical data from a connected client or server. The technique uses a combination of dynamic binary instrumentation, code coverage analysis, potential erroneous data generation and results analysis. The author also presents a new way to assess effectiveness of each test iteration using code coverage visualization. In the second part the author describes a technique implementation as a software framework that al-lows detecting vulnerabilities in network and file – based applications with high-level of visualization and testing automatiza-tion. In addition, the author conducted a large-scale experimental evaluation of the system on 17 different network applications for different operation systems. The results of the experiment confirmed that system suit s well for vulnerability detection in modern applications. Moreover, the experiments helped detect several previously unknown vulnerabilities in the popular and wide-spread applications.

13. Ability measuring researches of a fpga hardware module for load testing tasks [№4 за 2014 год]
Author: Borodin A.A.
Visitors: 7699
Quality and reliability assurance of information system is an important task. Nowadays it is mostly solved using load testing. Constant progress of information technologies requires increasing testing efficiency. Load testing is a complicated process that includes many stages. Theory and practice analysis showed that researcher’s attention to test launching is not enough. Efficiency of this stage depends on loader program performance quality and computer characteristics. Practically, resources of a single computer are not enough to produce the required amount of load. That is why there are load -creating methods based on distributed computing which have disadvantages. These methods include cloud and cluster computing as well as a group of computers connected together using local network. This paper presents experimental results of hardware loader characteristics measuring, created by the author based on FPGA to increase load creation stage efficiency. This device allows load testing process without using additional computers. During experiments, maximum load capability of a created prototype has been determined. The results were compared with characteristics of existing computer systems.

14. Online chair: cloud technology in higher education [№4 за 2014 год]
Authors: Telnov V.P., Myshev A.V.
Visitors: 13008
Ensuring computer networks and software scalability is one of the urgent problems for modern higher education. In the context of network traffic rapid growth IT-specialists have to spend more and more time and resources for ensuring bandwidth. At the same time, universities IT-budgets often lag behind the rate of growth. Universities need cost-effective, reliable and technological ways to meet the growing information needs while controlling costs. Created software product “Online chair” is based on the “cloud computing” concept, takes full advantage of high-tech solutions and public resources, with possible independence on specific service providers and licensed software. The product is aimed at the higher education institutions with budgetary constraints in terms of the acquisition and support their own computing infrastructure, network equipment and software. For budget universities of the Russian Federation software “Online chair” is available at no cost. The article considers the main characteristics of cloud computing, a brief analytical review of existing solutions for a higher school. In the main part the article formulates requirements to “Online chair” software, presents the main components of the software product, key architectural, technological and design solutions in the notation of UML diagrams, shows the results of testing the software and sample of user interface.

15. Development of an information processing prototype in respect of its cost [№4 за 2014 год]
Authors: Skripachev V.O., Chulkin M.O.
Visitors: 11807
There is a need to process heterogeneous information in complex information technology systems operating. It is useful to introduce the concept of information-technological process (ITP) for processing certain types of information. The object-oriented programming mechanisms are used when implementing specific ITP for information processing. They are encapsulation, inheritance, and polymorphism. When implementing the ITP software part it is reasonable to create a software prototype – a preliminary implementation of the proposed new software product. The main purpose of the prototype creating is elimination of ambiguities on the early stages of the development process. To develop a prototype it is better to use a parallel software development model. In addition, a prototype creation allows estimating the software creating costs and taking management decisions. The authors applied an algorithmic model of the software development estimation – a COnstructive COst MOdel (COCOMO). The article considers the features of the COCOMO application and the ITP software prototype using baseline for assessing labor intensity and costs of software development as a function of the program size. The size is expressed in thousands of estimating code lines. KLOC is used (kilo lines of code) to indicate the scope of program lines.

16. A Web-service for knowledge base generation on the basis of conceptual models [№4 за 2014 год]
Authors: Yurin A.Yu., Dorodnykh N.O.
Visitors: 13525
The paper describes a web-service for automated creation of the rule-type knowledge bases based on conceptual models. The service provides an automated analysis of the XML structure of IBM Rational Rose files (that contain description of class models) and FreeMind with subsequent selection of concepts and relations. Selected concepts and relations are represented as ontology. In turn the ontological model can be used for automated generation of rule bases and their visual modeling using RVML (Rule Visual Modeling Language) notation. Obtained models are used for generation of CLIPS codes (C Language Production System). The article presents descriptions of web-service functions, the architecture, algorithms for analysis of conceptual models, and tables for models transformation support. The web-service can be considered as one of the modules of the knowledge management system. This module is designed for retrieval, structuring and formalization of knowledge in different problem domains. The approbation and testing of the web-service are carried out in development of knowledge bases of rule-based expert systems for investigation the technogenic safety of hard-to-reach water objects.

17. A software package for inductive formation of medical knowledge bases [№4 за 2014 год]
Author: Smagin S.V.
Visitors: 8530
The article provides the description of InForMedKB (Inductive Formation of Medical Knowledge Bases) software package.It allows creating training sets (consisting of clinical histories from various branches of medicine) and forming medical knowledge bases inductivly. These knowledge bases are presented in the form accepted in the medical literature and contain descriptions of diseases (from specified branches of medicine), as well as explanations of these knowledge bases. The developed software package implements training algorithm which solves classification and clustering problems in their new definitions w hich are presented as a special case for problem of estimating the parameters of a reliability model that affects the quality of the training algorithm. This learning algorithm is developed for the useful and easily interpretable mathematical reliability model with parameters. It is a near real-life ontology of medical diagnostics (defined by a system of logical relationships with parameters). Presented algorithm finds parameter values for given model (medical knowledge base), which are close to the values that characterize the object domain of medical diagnostics. Using this software package and real data training set (containing clinical histories from "Acute abdomen" branch of medicine), a medical knowledge base is inductively formed. It has a high level of interpretability for a practicing physician. Descriptions of diseases included in the inductively formed knowledge base (expert evaluation) correspond to knowledge from scientific and academic medical literature. And sometimes they add descriptions of clinical implications dynamics. The formal representation of medical knowledge bases obtained using the software package allows using them in medical diagnostics expert systems.

18. Adaptive fuzzy systems on FOREL class taxonomy [№4 за 2014 год]
Authors: V.P. Shchokin, Chernyi, S.G., A.S. Bordug
Visitors: 10932
The article presents the results of developing a method of neuro-fuzzy self-organization structures in intelligent proc-ess control systems. The proposed modification of the basic algorithm can improve the control performance index of intelligent automated control systems at a reduced volume of calculations and corresponding increase in system performance. In classical train-ing rule fuzzy neural networks analyze the number of production rules, membership functions type, fuzzy inference algorithm type, etc. In the case of incorrect choice of these parameters fuzzy neural networks can be ineffective in the automation field. Th e devel-oped algorithm operation is based on the theory of sampling frequency and training frequency distribution. In the theory of control systems with discrete time it is determined that the sampling time T is usually selected according to the following rule of t humb: the value must exceed the maximum frequency of the system. In traditional adaptive control systems parameters are adjusted once e very sampling period, thus the sampling rate and update rate are not separated. We propose an expert method determining the concentra-tion coefficient of membership functions and sampling limits for further adjustments to the base of adaptive -established rules in or-der to reduce the algorithm running time and improve its efficiency when performing parametric synthesis of asymptotically stable intelligent control systems. According to the developed technique it is possible to compare the membership function parameter s which have been obtained as a result of the work of the modified adaptive algorithm of the Wang-Mendel fuzzy network, and the doubling parameters obtained from statistical processing of information systems solutions of the dynamic object state identif ication. We developed a systems operation algorithm taking into account the developed method of self-organization of neuro-fuzzy struc-tures based on the FOREL class taxonomy algorithm.

19. Genetic algorithm implementation for effective document subject search [№4 за 2014 год]
Authors: Ivanov V.K., Meskin P.I.
Visitors: 11919
The quality of documentary subject search or search for documents containing specifically coordinated information on a target subject is not always satisfactory. Despite the availability of powerful search engines for the Inter net information resources or special databases, the process remains time-consuming and poorly supported by software and methodologically. This paper describes the software implementation of a genetic algorithm for identifying and selecting most relevant results received during sequentially executed subject search operations. Simulated evolutionary process generates sustain able and effective population of search queries, forms search pattern of documents or semantic core, creates relevant sets of required documents, allows automatic classification of search results. The paper discusses the features of subject search, justifies the use of a genetic algorithm, describes arguments of the fitness function and describes basic steps and parameters of the algorithm. It also notes that the objective function or quality criteria is determined by the document position in search results built by the search engine for maximum number of different queries and semantic similarity of documents search pattern on a given subject. Software implementation is described in detail: general object models, user interface, the algorithm main library, morphological analysis modules, texts similarity analysis modules, search modules, database management modules, metadata management modules. The information on module classes composition and components is provided. The paper describes genetic algorithm software implementation that is one of the elements of Intelligent Distributed Information Management System for Innovations in Science and Education powered by the Russian Foundation of Basic Research. The algorithm plays an important role in functioning of the adaptive search engines. It is noted that developed algorithm software creates a sufficiently broad basis for further research and development.

20. Automatic feature selection system for human emotion recognition in speech communication [№4 за 2014 год]
Authors: Brester C.Yu., Semenkin E.S., Sidorov M.Yu.
Visitors: 7761
During the human-machine communication a number of problems related to voice processing should be solved. In addition to the speech recognition problem, there are several important issues such as a speaker, gender or age identification and speech-based emotion recognition. The amount of acoustic characteristics extracted from the signal is tremendously high (hundreds or even thousands): features may correlate with each other, contain noisy data or have low variation level that decrease the accuracy of involved classifiers. Therefore it is vitally important to select informative features automatically during the recognition process. This paper considers two feature selection techniques. Both of them are based on using the self-adaptive multi-objective genetic algorithm that is adjusted while the problem is being solved. The main advantages of this heuristic optimization procedure are the simplicity of coding the informative feature subsystem and the opportunity to optimize both discrete and continuous criteria. The probabilistic neural network is used as a classifier. Effectiveness investigation of the developed approaches was conducted on the set of emotion recognition problems: data bases contained speech signals in English and German languages. During the experiments it was revealed that application of the described feature selection procedures might lead to increasing of the classification accuracy (relative improvement was by up to 22,7 %). Moreover, it became possible to reduce the dimension of the feature vector significantly (from 384 to 64,8 attributes at the average). The proposed schemes demonstrate higher effectiveness compared with Principal Component Analysis. The described methods might be applied for solving the speaker identification problem, recognizing speaker’s gender, age or other personal characteristics that also implies the opportunity to use them as the algorithmic core in the intellectual modules of dialogue systems.

← Preview | 1 | 2 | 3 | 4 | 5 | Next →