All issues

The article is devoted to first-order adaptive methods for optimization problems with relatively strongly convex functionals. The concept of relatively strong convexity significantly extends the classical concept of convexity by replacing the Euclidean norm in the definition by the distance in a more general sense (more precisely, by Bregman’s divergence). An important feature of the considered classes of problems is the reduced requirements concerting the level of smoothness of objective functionals. More precisely, we consider relatively smooth and relatively Lipschitz-continuous objective functionals, which allows us to apply the proposed techniques for solving many applied problems, such as the intersection of the ellipsoids problem (IEP), the Support Vector Machine (SVM) for a binary classification problem, etc. If the objective functional is convex, the condition of relatively strong convexity can be satisfied using the problem regularization. In this work, we propose adaptive gradient-type methods for optimization problems with relatively strongly convex and relatively Lipschitzcontinuous functionals for the first time. Further, we propose universal methods for relatively strongly convex optimization problems. This technique is based on introducing an artificial inaccuracy into the optimization model, so the proposed methods can be applied both to the case of relatively smooth and relatively Lipschitz-continuous functionals. Additionally, we demonstrate the optimality of the proposed universal gradient-type methods up to the multiplication by a constant for both classes of relatively strongly convex problems. Also, we show how to apply the technique of restarts of the mirror descent algorithm to solve relatively Lipschitz-continuous optimization problems. Moreover, we prove the optimal estimate of the rate of convergence of such a technique. Also, we present the results of numerical experiments to compare the performance of the proposed methods.

The paper describes a method for recognizing authors of literary texts by the proximity of fragments into which a separate text is divided to the standard of the author. The standard is the empirical frequency distribution of letter combinations, built on a training sample, which included expertly selected reliably known works of this author. A set of standards of different authors forms a library, within which the problem of identifying the author of an unknown text is solved. The proximity between texts is understood in the sense of the norm in L1 for the frequency vector of letter combinations, which is constructed for each fragment and for the text as a whole. The author of an unknown text is assigned the one whose standard is most often chosen as the closest for the set of fragments into which the text is divided. The length of the fragment is optimized based on the principle of the maximum difference in distances from fragments to standards in the problem of recognition of «friend–foe». The method was tested on the corpus of domestic and foreign (translated) authors. 1783 texts of 100 authors with a total volume of about 700 million characters were collected. In order to exclude the bias in the selection of authors, authors whose surnames began with the same letter were considered. In particular, for the letter L, the identification error was 12%. Along with a fairly high accuracy, this method has another important property: it allows you to estimate the probability that the standard of the author of the text in question is missing in the library. This probability can be estimated based on the results of the statistics of the nearest standards for small fragments of text. The paper also examines statistical digital portraits of writers: these are joint empirical distributions of the probability that a certain proportion of the text is identified at a given level of trust. The practical importance of these statistics is that the carriers of the corresponding distributions practically do not overlap for their own and other people’s standards, which makes it possible to recognize the reference distribution of letter combinations at a high level of confidence.

Data hiding in digital images is a promising direction of cybersecurity. Digital steganography methods provide imperceptible transmission of secret data over an open communication channel. The information embedding efficiency depends on the embedding imperceptibility, capacity, and robustness. These quality criteria are mutually inverse, and the improvement of one indicator usually leads to the deterioration of the others. A balance between them can be achieved using metaheuristic optimization. Metaheuristics are a class of optimization algorithms that find an optimal, or close to an optimal solution for a variety of problems, including those that are difficult to formalize, by simulating various natural processes, for example, the evolution of species or the behavior of animals. In this study, we propose an approach to data hiding in the hybrid spatial-frequency domain of digital images based on metaheuristic optimization. Changing a block of image pixels according to some change matrix is considered as an embedding operation. We select the change matrix adaptively for each block using metaheuristic optimization algorithms. In this study, we compare the performance of three metaheuristics such as genetic algorithm, particle swarm optimization, and differential evolution to find the best change matrix. Experimental results showed that the proposed approach provides high imperceptibility of embedding, high capacity, and error-free extraction of embedded information. At the same time, storage of change matrices for each block is not required for further data extraction. This improves user experience and reduces the chance of an attacker discovering the steganographic attachment. Metaheuristics provided an increase in imperceptibility indicator, estimated by the PSNR metric, and the capacity of the previous algorithm for embedding information into the coefficients of the discrete cosine transform using the QIM method [Evsutin, Melman, Meshcheryakov, 2021] by 26.02% and 30.18%, respectively, for the genetic algorithm, 26.01% and 19.39% for particle swarm optimization, 27.30% and 28.73% for differential evolution.

Distributed saddle point problems (SPPs) have numerous applications in optimization, matrix games and machine learning. For example, the training of generated adversarial networks is represented as a min-max optimization problem, and training regularized linear models can be reformulated as an SPP as well. This paper studies distributed nonsmooth SPPs with Lipschitz-continuous objective functions. The objective function is represented as a sum of several components that are distributed between groups of computational nodes. The nodes, or agents, exchange information through some communication network that may be centralized or decentralized. A centralized network has a universal information aggregator (a server, or master node) that directly communicates to each of the agents and therefore can coordinate the optimization process. In a decentralized network, all the nodes are equal, the server node is not present, and each agent only communicates to its immediate neighbors.

We assume that each of the nodes locally holds its objective and can compute its value at given points, i. e. has access to zero-order oracle. Zero-order information is used when the gradient of the function is costly, not possible to compute or when the function is not differentiable. For example, in reinforcement learning one needs to generate a trajectory to evaluate the current policy. This policy evaluation process can be interpreted as the computation of the function value. We propose an approach that uses a smoothing technique, i. e., applies a first-order method to the smoothed version of the initial function. It can be shown that the stochastic gradient of the smoothed function can be viewed as a random two-point gradient approximation of the initial function. Smoothing approaches have been studied for distributed zero-order minimization, and our paper generalizes the smoothing technique on SPPs.

The paper considers a method for studying panel data based on the use of agglomerative hierarchical clustering — grouping objects based on the similarities and differences in their features into a hierarchy of clusters nested into each other. We used 2 alternative methods for calculating Euclidean distances between objects — the distance between the values averaged over observation interval, and the distance using data for all considered years. Three alternative methods for calculating the distances between clusters were compared. In the first case, the distance between the nearest elements from two clusters is considered to be distance between these clusters, in the second — the average over pairs of elements, in the third — the distance between the most distant elements. The efficiency of using two clustering quality indices, the Dunn and Silhouette index, was studied to select the optimal number of clusters and evaluate the statistical significance of the obtained solutions. The method of assessing statistical reliability of cluster structure consisted in comparing the quality of clustering on a real sample with the quality of clustering on artificially generated samples of panel data with the same number of objects, features and lengths of time series. Generation was made from a fixed probability distribution. At the same time, simulation methods imitating Gaussian white noise and random walk were used. Calculations with the Silhouette index showed that a random walk is characterized not only by spurious regression, but also by “spurious clustering”. Clustering was considered reliable for a given number of selected clusters if the index value on the real sample turned out to be greater than the value of the 95% quantile for artificial data. A set of time series of indicators characterizing production in the regions of the Russian Federation was used as a sample of real data. For these data only Silhouette shows reliable clustering at the level p < 0.05. Calculations also showed that index values for real data are generally closer to values for random walks than for white noise, but it have significant differences from both. Since three-dimensional feature space is used, the quality of clustering was also evaluated visually. Visually, one can distinguish clusters of points located close to each other, also distinguished as clusters by the applied hierarchical clustering algorithm.

Non-smooth optimization often arises in many applied problems. The issues of developing efficient computational procedures for such problems in high-dimensional spaces are very topical. First-order methods (subgradient methods) are well applicable here, but in fairly general situations they lead to low speed guarantees for large-scale problems. One of the approaches to this type of problem can be to identify a subclass of non-smooth problems that allow relatively optimistic results on the rate of convergence. For example, one of the options for additional assumptions can be the condition of a sharp minimum, proposed in the late 1960s by B. T. Polyak. In the case of the availability of information about the minimal value of the function for Lipschitz-continuous problems with a sharp minimum, it turned out to be possible to propose a subgradient method with a Polyak step-size, which guarantees a linear rate of convergence in the argument. This approach made it possible to cover a number of important applied problems (for example, the problem of projecting onto a convex compact set). However, both the condition of the availability of the minimal value of the function and the condition of a sharp minimum itself look rather restrictive. In this regard, in this paper, we propose a generalized condition for a sharp minimum, somewhat similar to the inexact oracle proposed recently by Devolder – Glineur – Nesterov. The proposed approach makes it possible to extend the class of applicability of subgradient methods with the Polyak step-size, to the situation of inexact information about the value of the minimum, as well as the unknown Lipschitz constant of the objective function. Moreover, the use of local analogs of the global characteristics of the objective function makes it possible to apply the results of this type to wider classes of problems. We show the possibility of applying the proposed approach to strongly convex nonsmooth problems, also, we make an experimental comparison with the known optimal subgradient method for such a class of problems. Moreover, there were obtained some results connected to the applicability of the proposed technique to some types of problems with convexity relaxations: the recently proposed notion of weak $\beta$-quasi-convexity and ordinary quasiconvexity. Also in the paper, we study a generalization of the described technique to the situation with the assumption that the $\delta$-subgradient of the objective function is available instead of the usual subgradient. For one of the considered methods, conditions are found under which, in practice, it is possible to escape the projection of the considered iterative sequence onto the feasible set of the problem.

This paper considers the problem of water consumption in the regions of Russia with low water availability. We provide a review of the existing methods to control quality and quantity of water resources at different scales — from households to worldwide. The paper itself considers regions with low “water availability” parameter which is amount of water per person per year. Special attention is paid to the regions, where this parameter is low because of natural features of the region, not because of high population. In such regions many resources are spend on water processing infrastructure to store water and transport water from other regions. In such regions the main water consumers are industry and agriculture.

We propose dynamic two-level hierarchical model which matches water consumption of a region with its gross regional product. On the top level there is a regional administration (supervisor) and on the lower level there are region enterprises (agents). The supervisor sets fees for water consumption. We study the model with Pontryagin’s maximum principle and provide agents’s optimal control in analytical form. For the supervisor’s control we provide numerical algorithm. The model has six free coefficients, which can be chosen so the model represents a particular region. We use data from Russia Federal State Statistics Service for identification process of a model. For numerical analysis we use trust region reflective algorithms. We provide calculations for a few regions with low water availability. It is shown that it is possible to reduce water consumption of a region more than by 20% while gross regional product drop is less than 10%.

The paper contains a mathematical model for the optimal location of enterprises producing fuel from renewable wood waste for the regional distributed heating supply system. Optimization is based on total cost minimization of the end product – the thermal energy from wood fuel. A method for solving the problem is based on genetic algorithm. The paper also shows the practical results of the model by example of Udmurt Republic.

The probabilistic topic model of a text document collection finds two matrices: a matrix of conditional probabilities of topics in documents and a matrix of conditional probabilities of words in topics. Each document is represented by a multiset of words also called the “bag of words”, thus assuming that the order of words is not important for revealing the latent topics of the document. Under this assumption, the problem is reduced to a low-rank non-negative matrix factorization governed by likelihood maximization. In general, this problem is ill-posed having an infinite set of solutions. In order to regularize the solution, a weighted sum of optimization criteria is added to the log-likelihood. When modeling large text collections, storing the first matrix seems to be impractical, since its size is proportional to the number of documents in the collection. At the same time, the topical vector representation (embedding) of documents is necessary for solving many text analysis tasks, such as information retrieval, clustering, classification, and summarization of texts. In practice, the topical embedding is calculated for a document “on-the-fly”, which may require dozens of iterations over all the words of the document. In this paper, we propose a way to calculate a topical embedding quickly, by one pass over document words. For this, an additional constraint is introduced into the model in the form of an equation, which calculates the first matrix from the second one in linear time. Although formally this constraint is not an optimization criterion, in fact it plays the role of a regularizer and can be used in combination with other regularizers within the additive regularization framework ARTM. Experiments on three text collections have shown that the proposed method improves the model in terms of sparseness, difference, logLift and coherence measures of topic quality. The open source libraries BigARTM and TopicNet were used for the experiments.

In the first part of the paper we present convergence analysis of AGMsDR method on a new class of functions — in general non-convex with $M$-Lipschitz-continuous gradients that satisfy Polyak – Lojasiewicz condition. Method does not need the value of $\mu^{PL}>0$ in the condition and converges linearly with a scale factor $\left(1 - \frac{\mu^{PL}}{M}\right)$. It was previously proved that method converges as $O\left(\frac1{k^2}\right)$ if a function is convex and has $M$-Lipschitz-continuous gradient and converges linearly with a~scale factor $\left(1 - \sqrt{\frac{\mu^{SC}}{M}}\right)$ if the value of strong convexity parameter $\mu^{SC}>0$ is known. The novelty is that one can save linear convergence if $\frac{\mu^{PL}}{\mu^{SC}}$ is not known, but without square root in the scale factor.

The second part presents modification of AGMsDR method for solving problems that allow alternating minimization (Alternating AGMsDR). The similar results are proved.

As the result, we present adaptive accelerated methods that converge as $O\left(\min\left\lbrace\frac{M}{k^2},\,\left(1-{\frac{\mu^{PL}}{M}}\right)^{(k-1)}\right\rbrace\right)$ on a class of convex functions with $M$-Lipschitz-continuous gradient that satisfy Polyak – Lojasiewicz condition. Algorithms do not need values of $M$ and $\mu^{PL}$. If Polyak – Lojasiewicz condition does not hold, the convergence is $O\left(\frac1{k^2}\right)$, but no tuning needed.

We also consider the adaptive catalyst envelope of non-accelerated gradient methods. The envelope allows acceleration up to $O\left(\frac1{k^2}\right)$. We present numerical comparison of non-accelerated adaptive gradient descent which is accelerated using adaptive catalyst envelope with AGMsDR, Alternating AGMsDR, APDAGD (Adaptive Primal-Dual Accelerated Gradient Descent) and Sinkhorn's algorithm on the problem dual to the optimal transport problem.

Conducted experiments show faster convergence of alternating AGMsDR in comparison with described catalyst approach and AGMsDR, despite the same asymptotic rate $O\left(\frac1{k^2}\right)$. Such behavior can be explained by linear convergence of AGMsDR method and was tested on quadratic functions. Alternating AGMsDR demonstrated better performance in comparison with AGMsDR.