|
«INFORMATICS AND APPLICATIONS»
Scientific journal
Volume 20, Issue 3, 2026
Content | About Authors
Abstract and Keywords
- I. V. Chizhov M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: In 2023, an approach to attacking the McEliece cryptosystem was proposed that is based on the study of so-called quadratic relationship codes, which are closely connected with the Schur-Hadamard product. These codes are constructed from quadratic forms that vanish on the columns of a generator matrix of a linear code. The effectiveness of this approach was recently demonstrated by an attack on the McEliece cryptosystem built upon binary Goppa codes of small degree. This attack exploits the fact that the quadratic relationship code contains a quadratic form of a relatively small rank. In the present paper, a systematic study is carried out of linear codes whose quadratic relationship code contains forms of rank 2 and lower. A special case is considered where the quadratic relationship code contains a reducible quadratic form, i. e., a form that decomposes into a product of two nonzero linear forms. Finally, the decoding problem is addressed for codes whose quadratic relationship codes contain no reducible quadratic forms or contain relatively few of them.
Keywords: code of quadratic relationships; Schur-Hadamard square of a code; McEliece cryptosystem; minimal code; decoding problem
- A. V. Borisov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation, M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation
- Yu. N. Kurinov M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation
Abstract: The first part of this series presents the theoretical foundations required for the proper formulation and solution of a finite-horizon stochastic control problem under incomplete information. The controlled system is represented by a class of Markov jump processes (MJPs) with a finite state space. The optimality criterion is defined as the expected value of an integral loss functional. The observations available for control synthesis are functions of the controlled state, measured without external noise but subject to a time delay. The class of admissible controls consists of processes that are predictable with respect to the natural filtration generated by the observations and that satisfy a set of geometric and integral constraints. A universal filtered probability space is constructed to provide a solution to the martingale problem associated with this class of controlled MJPs. In addition, conditions ensuring the continuity of both the MJP state trajectory and the objective functional with respect to the applied control are established.
Keywords: controlled Markov jump process; Wiener-Poisson space; martingale problem; Poisson stochastic measure
- A. V. Bosov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- S. A. Bosov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- I. V. Uryupin Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: New computationally stable algorithms are proposed for adaptive filtering of the state of an object moving in an aquatic environment, based on acoustic observations with random time delays. A previously developed version of the extended Kalman filter, adapted to the time-delay model and built upon the method of linear pseudoobservations, is supplemented with two adaptive filter modifications for the case of unknown covariances of dynamic disturbances and measurement errors. To estimate the covariances, stable adaptive Kalman filter schemes are used, derived from the analysis of measurement residuals with respect to the filtering estimates. These schemes are improved by additional optimization in order to account for the positive diagonal structures of the unknown covariances assumed by the motion model and generated by the linear pseudoobservations. A large-scale numerical experiment was carried out, confirming the operability of the proposed adaptation schemes. The calculations employed the same model as in previous works, which made it possible to separately assess the influence of the adaptive formulation and to outline possible directions for further development of the methodology.
Keywords: stochastic filtering; stochastic system with random observation delays; extended Kalman filter (EKF); EKF based on the method of linear pseudoobservations; adaptive Kalman filter (AKF)
- E. I. Melezhnikov Department of Mathematical Statistics, Faculty of Computational Mathematics and Cybernetics, M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation, Moscow Center for Fundamental and Applied Mathematics, M.V. Lomonosov Moscow State University, 1 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation
- O. V. Shestakov Department of Mathematical Statistics, Faculty of Computational Mathematics and Cybernetics, M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation, Moscow Center for Fundamental and Applied Mathematics, M.V. Lomonosov Moscow State University, 1 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The problem of threshold processing of a sparse signal with noisy observations is considered. The number of observations is assumed to be random and generated by a Poisson process with a given intensity. The behavior of the mean-square risk and its estimate based on Stein's unbiased risk estimate (SURE) is investigated. Special attention is paid to the effect of randomness in the sample size on signal recovery accuracy and on the properties of risk estimation. An upper bound for the risk under optimal threshold selection is obtained and it is shown that its asymptotic order coincides with that in the deterministic case. In addition, a central limit theorem and a strong law of large numbers for the SURE risk estimate are proved. Thus, the stability of the asymptotic properties of thresholding procedures is established when passing to a model defined by a Poisson process. The obtained results are applicable to problems of streaming data analysis, where observations arrive at random time moments and the sample size is not fixed in advance. This extends the applicability of thresholding methods to statistical models with a random number of observations.
Keywords: thresholding; mean square risk; risk estimate; sparse model; Poisson process; central limit theorem; strong law of large numbers
- Yu. E. Malashenko Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- I. A. Nazarova Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- M. V. Kozlov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: Within computational experiments, monopolistic and mixed routing strategies for outgoing nodal multiflows in multiuser network systems are studied. Two basic monopolistic dispatching schemes for internodal flows are analyzed: routing along shortest paths and routing along maximum flow transmission paths. For each strategy, normalized vectors of outgoing multiflows - transmitted in a monopolistic mode from every source node to all destination nodes - are formulated. The characteristic vectors of edge loading are determined when normalized multiflows from source vertices pass through the network. To evaluate network performance under mixed dispatching strategies, convex combinations of these characteristic vectors are constructed. For all characteristic vectors, the standard deviations of component values, the relative coefficients of variation, and the magnitude of the maximum excess of the average edge loading level are calculated. A comparative analysis of the aggregated functional indicators for each node and the proposed routing methods is carried out. The impact of nodal multiflows on overall network load is investigated and potential overload points and bottlenecks are identified.
The experimental results are illustrated using specialized diagrams.
Keywords: multicommodity flow model; outgoing nodal multiflow; network load distribution
- M. P. Krivenko Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: Equivalent definitions of ellipsoids are considered as well as formulations of analytic geometry problems aimed at interpreting the cluster structure of data. The corresponding solution algorithms are mathematically substantiated. Two examples from the field of data analysis are discussed based on a Gaussian mixture model. Ellipsoids are used to interpret the data clusters corresponding to the elements of the mixture. The first example focuses on modeling reference values by describing the empirical distribution of multivariate patient data, including age and PSA (Prostate-Specific Antigen) biomarker measurements. The proposed mixture-based solutions demonstrate clear advantages and enable the direct application of ellipse-based visualization methods to identify specific features of the object under study. The second example considers a consolidated approach for analyzing longitudinal data when a series of multidimensional object characteristics is represented as a single vector of observed values. To demonstrate the emerging capabilities of data analysis, the problem of early diagnosis of cancer using PSA biomarkers is investigated. The advantage of the consolidating method is confirmed by a high degree of separability of sets of cluster elements of the mixture measured by the number of pairwise intersections of the corresponding ellipsoids. Analysis of the degree of intersection of sets of ellipsoids for different classes of diagnoses can further help identify the contribution of individual data clusters to classification errors.
Keywords: mixture of normal distributions; representation of ellipsoids; ellipsoid processing algorithms; reference values; serial data classification; longitudinal analysis; consolidation approach\
- A. M. Dostovalova Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- A. K. Gorshenin Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The paper proposes a method for solving the continuous incremental classification problem, which consists in discovering new classes in data streams under the condition of absent labeling for limited time series and tabular data. To generate accurate pseudolabels for new data that complement previously identified patterns, a specialized neural network was developed whose architecture is informed by a deep mixture ofGaussian distributions and which implements learning based on a contrastive loss function. The method was tested on publicly available datasets using various discriminator architectures, including the Transformer. The performance of the proposed method was compared against uninformed networks and machine learning methods for pseudo-labeling.
The informed network demonstrates superior generalization capabilities in detecting new classes within unlabeled data, particularly under conditions of small-scale training datasets. The gain in the harmonic Ff metric, which balances the recognition accuracy for objects of old and new classes, reaches 66.32% (with an average value of 15.28%), while the macroaveraged classification accuracy for the dataset based on the Ff"8 metric increases by 52.86% (with an average value of 11.8%).
Keywords: incremental learning; probability informing; new classes discovery; unlabeled data; deep Gaussian mixture models
- I. Yu. Torshin Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: In the first part of this work, the main theoretical directions for generating problem-oriented metrics on sets of objects (pQ-metrics) using metrics on features (pL-metrics) were systematized. In this paper, criteria for tuning PQ-metrics are obtained using concepts of compactness. Computational experiments were conducted on 1000 independent data samples with 5000 pQ-metrics synthesized according to 20 proposed approaches for numerical forecasting by fc-nearest neighbor algorithms. Computational experiments demonstrated that in 95% cases, pQ-metrics of three types were effective: (i) those based on synthetic numerical features (61%); (ii) vector- based with pairwise matching (22%); and (iii) noncommutative metrics based on pL-distance arrays (12%). Importantly, pQ-metrics based on synthetic numerical features provided the best result for 83% of the datasets, which makes this approach the most promising.
Keywords: topological data analysis; distance functions on objects; algebraic approach; feature value analysis theory
- S. V. Listopad Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The article develops a conceptual framework for the multiaspect identification of problem structures with the aim of constructing automated problem-solving methods based on hybrid intelligent multiagent systems.
The topic is relevant due to the weak formalization, heterogeneity, and complexity of practical problems, which require significant effort from developers of artificial intelligence systems. The paper summarizes the most well- known bases for problem decomposition, demonstrates the need for their combined application, and proposes a formalized model for automatic multiaspect problem identification, ensuring the selection of decomposition bases in accordance with the types of uncertainty characteristic of the task.
Keywords: problem; conceptual model; decomposition; reduction; team of specialists; hybrid intelligent multiagent system
- A. A. Grusho Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- N. A. Grusho Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- M. I. Zabezhailo Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- A. A. Zatsarinny Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- V. O. Piskovski Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The application of network-centric technologies in control and decision support systems is determined by the concept of development of monitoring of peripheral nodes of the network-centric system (NCS) as well as the conditions for ensuring information security. The work is devoted to one class of attacks on such systems, which uses hidden distortion of the properties of information necessary for proper monitoring and control. The main goal of the Center in the NCS is described as the main controlling cause, consisting of properties that ultimately generate controlling effects throughout the NCS. A mathematical model of the control actions in the NCS is constructed using directed acyclic graphs. Examples of attacks on the NCS, the possibility of their identification, and countering such attacks are considered. It is demonstrated how the prevention of the development of an attack can be provided by identifying the missing consequences of the correct property as well as the use of deceptive properties.
Keywords: information security; network-centric systems; anomalies; detection of anomalies
|
