«Systems and Means of Informatics»
Scientific journal
Volume 34, Issue 1, 2024
Content | About Authors
Abstract and Keywords
Abstract: The paper is devoted to nonlinear correlation methods for analytical modeling in differential stochastic systems with unsolved derivatives (StS USD) and random parameters. Survey is given. Necessary notations concerning integral canonical expansions (ICE) and its linear and nonlinear transforms are presented. It is shown how differential StS USD can be reduced to differential StS. Basic quality analysis algorithms for reducible StS USD are described. Special attention is paid to multicomponent ICE theory of stochastic processes and StS USD reducible to the differential ones. Two types of nonlinear transforms based on linear ICE regression are developed. Normal approximation method is used for ordinary differential equations for conditional probabilistic characteristics: mathematical expectations, covariance matrix, and matrix of covariance functions. For uncondional characteristics, ICE method is implemented. Analytical modeling methods are presented both for stationary and nonstationary regimes. An illustrative example is given. Directions of future generalizations are given.
Keywords: analytical modeling; integral canonical expansion (ICE); normal approximation method (NAM); stochastic systems with unsolved derivatives (StS USD); conditional and unconditional characteristics
- I. A. Kochetkova
- A. S. Vlaskina
- P. L. E. Eseng Mangue
- V. S. Shorgin
Abstract: Elastic resource management is anticipated to be prevalent in the forthcoming 6G, or 2030, networks. This approach entails not only adaptive allocation of resources but also the consideration of evolving service requirements, user behavior, and the dynamic state of the radio channel. One of the methods for distributing resources among network slices may not be equitable in terms of the required data transfer rates but aims for fairness concerning the data transmission delays. The paper formalizes the issue of resource allocation based on data transmission latency as a queuing system with a discriminatory processor-sharing discipline for classes of elastic traffic. The authors propose determining the discipline parameters - the weights for each class - according to the residual data transmission time. The fairness of resource allocation is evaluated using the Jain's fairness index which is based on average transmission times.
Keywords: wireless network; network 2030; 6G; resource allocation; latency fairness; elastic traffic; queuing system; processor sharing
- Yu. A. Stepchenkov
- D. V. Khilko
- Yu. G. Diachenko
- N. V. Morozov
- D. Yu. Stepchenkov
- G. A. Orlov
Abstract: Self-timed (ST) digital circuits have undoubted advantages over synchronous counterparts. However, ST circuit synthesis currently requires the user to have deep knowledge in the ST circuit technology field, since its automation level is still far from the level achieved in industrial computer- aided design systems for synchronous circuits and is focused on assigning the synthesized circuit behavior in specific formats. Designing of ST circuits is more labor-intensive and specific in comparison with synchronous circuits because of the need to adhere to the strict principles for their implementation. Desynchronization is an important stage in the ST circuit synthesis based on the original Verilog description of the circuit operation algorithm. It provides circuit separation from the global clock and asynchronous request-acknowledge interaction usage preparation. The article considers the desynchronization implementation methodology and its formalization principles. The proposed method ensures an ST circuit correct construction based on heuristic algorithms determining the relationships between functional blocks in the synthesized circuit and organizing their interaction in strict accordance with the ST circuit operation discipline.
Keywords: synchronous circuit; Verilog description; self-timed circuit; automated synthesis; desynchronization; indication; control
- D. V. Fedoseev
- A. S. Stepanov
- L. V. Illarionova
- V. A. Eliseev
- I. O. Prokhorets
- A. N. Frolov
- A. I. Shamurina
Abstract: Digitalization of agriculture is one of the strategic directions of economic development. To store and process the results of satellite monitoring of agricultural lands, it is planned to develop information systems (IS) at the federal and regional levels. The article discusses the issues of developing an appropriate IS for the southern part of the Far East. A two-level system including client and server levels is described. The implemented database (DB) management system based on PostgreSQL with the PostGIS extension has a cluster structure where the subject of the Russian Federation is considered as a cluster. The cluster is logically divided into 5 main modules which are grouped tables with data of different types. Triggers and trigger functions are used to optimize the operation of the IS and the possibility of sharing DB is also provided. The DB is being tested and filled for the Khabarovsk and the Amur Regions.
Keywords: information systems; database; satellite monitoring; agriculture
- I. O. Prokhorets
- A. S. Stepanov
Abstract: Automated classification of arable lands using machine learning methods is one of the most important tasks in the transition to digital agriculture. The classification of arable lands in the Khabarovsk Region was carried out using random forest (RF), minimum distance (MD), and K-means clustering methods based on Sentinel-2 images for July, August, September, and October 2022. The values of spectral bands, EVI (Enhanced Vegetation Index), and NDVI (Normalized Difference Vegetation Index) were considered as input data. Based on the results of statistical processing, it was found that the RF method demonstrated the greatest stability when changing the date of shooting and the type of input data. The accuracy of recognition of arable lands in the Khabarovsk Region in 2022 was 92.5% when using NDVI values calculated from the September Sentinel-2 image in the classifier. The proposed approach can be used for automated classification and subsequent mapping with expert correction of arable lands in the southern part of the Far East.
Keywords: mapping; machine learning; satellite monitoring; arable lands; classification; Khabarovsk Region
Abstract:Today, astronomers are faced with a challenge of handling vast volume
of data as modern astronomical surveys are capable of generating terabytes of
data in a single night. One of such survey is the Zwicky Transient Facility
(ZTF), an automated sky survey that provides approximately a million alerts per
observational night.However, a significant part of the detected objects turn out to
be artifacts, i. e., phenomena of a nonastrophysical origin. Therefore, specialists
must invest time in manually classifying these objects as there is currently
no efficient method that can perform this task without human intervention.
The goal of the work is the development of an algorithm to predict whether
a light curve from the ZTF data releases (DRs) has a bogus nature or not,
based on the sequence of frames. A labeled dataset provided by experts from
SNAD team was utilized, comprising 2230 frames series. Due to substantial
size of the frame sequences, the application of a variational autoencoder (VAE)
was deemed necessary for mapping the images into lower-dimensional vectors.
For the task of binary classification based on sequences of compressed frame
vectors, a recurrent neural network (RNN) was employed. Several neural network
models were considered and the quality metrics were assessed using k-fold cross-
validation. The final performance metrics, including ROC-AUC = 0.86 ± 0.01
and Accuracy = 0.80 ± 0.02, suggest that the model has practical utility. The
code implementing the algorithm is available on GitHub.
Keywords: neural network; data analysis; real-bogus classification
Abstract: The article discusses the problems of ensuring the necessary level of information security of artificial intelligence systems which is becoming one of the key factors in their widespread use. Based on the analysis of data on committed cyberattacks on well-known machine learning systems (MLS), a complete set of stages of the life cycle of computer attacks on MLS was formed, target MLS vulnerabilities for each stage of the life cycle were determined, methods of attacks on MLS vulnerabilities were systematized, and approaches to ensuring information security of the system were outlined.
Keywords: information security; artificial intelligence systems; machine learning; types of information protection
Abstract: The paper proposes a method for developing negotiation tactics by artificial heterogeneous intelligent agents of reflexive-active systems designed for computer modeling of reasoning in long-term teams of specialists in solving practical problems. Agents of such systems reflexively model each other's rea-soning developing strategies and tactics for negotiating, while solving a problem, in accordance with the models of counterparties which ensures self-organization of agents in a strong sense without explicit centralized control. Thanks to these features, reflexive-active systems of artificial heterogeneous intelligent agents are able to take into account such properties of practical problems as heterogeneity, dynamism, and complexity similar to simulated teams of specialists.
Keywords: team of specialists; reflection; negotiations; hybrid intelligent multiagent system
Abstract: The term "cloud computing" has firmly entered the area of information technologies. The requirements to modern cloud platforms in a flexible application-centric infrastructure are most effectively implemented with software-defined cloud platforms. The most important infrastructure component of the software-defined cloud is the network which not only consolidates all infrastructure components of the platform but also provides the client access to cloud services. In the software-defined cloud, a software-defined network looks most naturally being able to provide the cloud infrastructure with the necessary flexibility and manageability. The article examines some issues of the software-defined cloud platforms paying special attention to a number of network aspects from the existing terminological uncertainty regarding the data center infrastructures to some management features of the software-defined cloud networks and the role of the networking in an application-centric infrastructure.
Keywords: cloud platform; software-defined cloud; software-defined networking
- V. I. Budzko
- V. I. Medennikov
Abstract: Issues of determining the effectiveness of training specialists at agricultural Russian universities using mathematical models for assessing their digital traces left on various Internet sites are considered. The increasing importance of these traces at the present stage of development of higher education digitalization has been determined. It is shown that teachers and students will have qualitatively new means for exchanging educational and scientific knowledge and ideas and their effective transfer to the economy will be ensured based on the integration mechanisms of the Internet representation of educational institutions with their competent organization. This will increase the efficiency of using the country's intellectual potential. The impact of the digital footprint on the competitiveness of educational organizations is shown. To calculate ratings based on digital traces, mathematical models that formalize the identified digital footprints were developed and methods were formed on their basis which were used to compare the effectiveness of the use of educational and scientific information resources by universities. The impact of the imposed sanctions against Russia on the digital footprint was assessed by comparing the ratings obtained in 2022 with the ratings calculated in 2016. The dynamics of changes in the volume and quality of resources presented on these sites are given.
Keywords: agricultural universities; digital footprint; ratings; assessment methodology
- I. M. Adamovich
- O. I. Volkov
Abstract: The article is devoted to the further development of a distributed technology of concrete historical investigation support based on the principles of crowdsourcing and focused on a wide range of users who are not professional historians and biographers. Development is carried out by including in the technology an entity resolution algorithm for nominative documents processing that performs collective resolution in which entities for matching links are determined jointly. This algorithm is a modification of the greedy agglomerative clustering algorithm. The article provides a detailed description of the approach underlying the algorithm and provides its high-level pseudocode. The analysis of its effectiveness on data with varying degrees of ambiguity of names is given and the degree of ambiguity of names of concrete historical data is estimated. The conclusion about the expediency of including the algorithm in the technology is made. The directions of further research on determining the configurable parameters of the algorithm are outlined.
Keywords: concrete historical investigation; distributed technology; entity resolution; greedy algorithm; relational similarity measure
Abstract: The main structural and spatiotemporal characteristics of the author's informatics-cybernetic hierarchical-network model of the self-controlling system of Humankind are presented. The stages of its systemic global evo-lution are compared with the classification of types of sign systems (TSS) by A.B. Solomonick. A high degree of correspondence has been established be-tween the content of the stages of complication of basic information technologies (BIT) and spatial types of civilizations (STC) in the course of global evolution with the elements of the TSS classification in threes: (i) BIT "Signal poses/sounds/movements"-STC "pre-pre-civilizations"-TSS "Natural sign systems;" (ii) BIT "Mimics/gestures"-STC "pre-civilizations"-TSS "Im-age systems;" (iii) BIT "Speech/language" - STC "protocivilizations"-TSS "Language systems; (iv) BIT "Writing/reading" - STC "Local civilizations" - TSS "Writing systems;" (v) BIT "Replication of texts/book printing" - STC "regional and subcontinental civilizations"-TSS "Formalized sign systems of the first order;" and (vi) BIT "Local computer" - STC "Planetary Civiliza-tion"-TSS "Formalized sign systems of the second order." The possibility of supplementing the classification of TSS by introducing the concepts of "For-malized sign systems of the third order" and "Formalized sign systems of the fourth order" as an analogue to the following stages in global evolution BIT "Telecommunications/networks"- STC "Civilization of Near-Earth Space" and "nano-BIT (AI-BIT)" - STC "Civilization of Intermediate Cosmos," respectively. It is also proposed to expand the meaningful interpretation of the elements of the TSS classification by introducing into it the corresponding spatiotemporal parameters of informatics-cybernetic model and also consider TSS as defining cultural characteristics of the corresponding STC. It is concluded that the sequence of emergence and increasing the level of abstraction of the TSC can be interpreted as another reflection of the processes of the systemic global evolution of Humanity.
Keywords: self-controlling system of Humankind; systemic global evolution; informatics-cybernetic model; basic information technologies; spatial types of civilizations; types of sign systems; semiotic classification by A. B. Solomonick
|
