All issues

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.

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 work solves the problem of establishing the dependence of the potential for spatial selection of useful and interfering signals according to the signal-to-interference ratio criterion on the positioning error of user equipment during beamforming by their location at a base station, equipped with an antenna array. Configurable simulation parameters include planar antenna array with a different number of antenna elements, movement trajectory, as well as the accuracy of user equipment location estimation using root mean square error of coordinate estimates. The model implements three algorithms for controlling the shape of the antenna radiation pattern: 1) controlling the beam direction for one maximum and one zero; 2) controlling the shape and width of the main beam; 3) adaptive beamforming. The simulation results showed, that the first algorithm is most effective, when the number of antenna array elements is no more than 5 and the positioning error is no more than 7 m, and the second algorithm is appropriate to employ, when the number of antenna array elements is more than 15 and the positioning error is more than 5 m. Adaptive beamforming is implemented using a training signal and provides optimal spatial selection of useful and interfering signals without device location data, but is characterized by high complexity of hardware implementation. Scripts of the developed models are available for verification. The results obtained can be used in the development of scientifically based recommendations for beam control in ultra-dense millimeter-wave radio access networks of the fifth and subsequent generations.

Detecting controversy in online discussions is critical for managing public relations, as it helps inform various processes from policymaking to business. This work aims to expand approaches to online controversy detection based on the expressed emotions. Controversy was defined as an online content phenomenon of provoking disagreements and conflict. This study builds upon prior semantic methods by analyzing estimates of emotional connotations of messages. Modern language models for emotion recognition and named entity recognition are explored as tools of controversy detection. The outputs of these models were aggregated by entity to estimate the entity’s emotional connotation. The emotional divergence score based on the dispersion of emotions was proposed to quantify controversy in user content. Then, entities with sufficiently high emotional divergence relative to the domain of discussions were selected as markers of controversy. A case study of Reddit data related to Sri-Lankan 2022 political crisis was conducted, showing the capabilities of emotional divergence score in controversy detection. A total of two datasets were collected with different methodologies: one aimed at collecting earlier messages and another aimed at collecting more recent ones. The collected data contained discussions of policy, public figures, organizations and locations tied to the crisis. When measured on manually annotated data samples, the proposed method achieved a recall value of 0.705 and a precision value close to 0.496 for the first dataset, while recall of 0.716 and precision of 0.436 were recorded for the second dataset. The main factors that limit the precision were found to be the quality of underlying models and false positives: highly discussed non-controversial markers. Lastly, it was identified that a study of regular emotional distribution of social media content may be helpful for improving controversy detection quality.

The problem is to automate the correction of LaTeX documents. Each document is represented as a parse tree. The modified Zhang-Shasha algorithm is used to construct a mapping of tree vertices of the original document to the tree vertices of the edited document, which corresponds to the minimum editing distance. Vertex to vertex maps form the training set, which is used to generate rules for automatic correction. The statistics of the applicability to the edited documents is collected for each rule. It is used for quality assessment and improvement of the rules.

The purpose of this study is to develop a methodology for change points detection in time series, including financial data. The theoretical basis of the study is based on the pieces of research devoted to the analysis of structural changes in financial markets, description of the proposed algorithms for detecting change points and peculiarities of building classical and deep machine learning models for solving this type of problems. The development of such tools is of interest to investors and other stakeholders, providing them with additional approaches to the effective analysis of financial markets and interpretation of available data.

To address the research objective, a neural network was trained. In the course of the study several ways of training sample formation were considered, differing in the nature of statistical parameters. In order to improve the quality of training and obtain more accurate results, a methodology for feature generation was developed for the formation of features that serve as input data for the neural network. These features, in turn, were derived from an analysis of mathematical expectations and standard deviations of time series data over specific intervals. The potential for combining these features to achieve more stable results is also under investigation.

The results of model experiments were analyzed to compare the effectiveness of the proposed model with other existing changepoint detection algorithms that have gained widespread usage in practical applications. A specially generated dataset, developed using proprietary methods, was utilized as both training and testing data. Furthermore, the model, trained on various features, was tested on daily data from the S&P 500 index to assess its effectiveness in a real financial context.

As the principles of the model’s operation are described, possibilities for its further improvement are considered, including the modernization of the proposed model’s structure, optimization of training data generation, and feature formation. Additionally, the authors are tasked with advancing existing concepts for real-time changepoint detection.