All issues

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.

This is a brief review of the subjects, and an impression of some talks, which were given at the Schools on modelling complex biological systems. Those Schools reflected a logical progress in this way of thinking in our country and provided a place for collective “brain-storming” inspired by prominent scientists of the last century, such as A. A. Lyapunov, N. V. Timofeeff-Ressovsky, A. M. Molchanov. At the Schools, general issues of methodology of mathematical modeling in biology and ecology were raised in the form of heated debates, the fundamental principles for how the structure of matter is organized and how complex biological systems function and evolve were discussed. The Schools served as an important sample of interdisciplinary actions by the scientists of distinct perceptions of the World, or distinct approaches and modes to reach the boundaries of the Unknown, rather than of different specializations. What was bringing together the mathematicians and biologists attending the Schools was the common understanding that the alliance should be fruitful. Reported in the issues of School proceedings, the presentations, discussions, and reflections have not yet lost their relevance so far and might serve as certain guidance for the new generation of scientists.

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.

This study proposes a methodology for anomaly detection in educational assessment data, demonstrated on the case of the 2023–2024 Basic State Exam (BSE) in mathematics in Russia. The relevance of the study is related to the absence of mandatory video surveillance during the examination period, which creates a risk of potential rule violations both by individual students and by entire educational institutions. By analyzing the distribution of primary scores, we identify a big spike in the area between grades 2 and 3 as a specific pattern in results that may indicate cases of cheating during the exam. To determine the most suspicious results, two anomaly criteria were constructed. The first criterion relies on comparing the magnitude of the spike in empirical distribution function in school’s results with the corresponding regional average level. This criterion made it possible to identify 47 educational institutions with abnormally high values of the spike. The second (general) criterion was derived from comparing students’ scores on the examination with their performance on a diagnostic mathematics test conducted in grade 8 under video surveillance. This comparison is appropriate because almost the same group of students took part in both assessments. This approach helps reduce the number of detected anomalies by distinguishing those more likely to reflect actual protocol violations from those arising due to the specific characteristics of a particular student population and their exam preparation within a given educational institution. The application of the oneclass support vector machine method enabled the identification of 12 schools with atypical anomalous results. The proposed methodology could be useful for the detection of potential cases of cheating during exams and the development of methods for preventing such behavior. In particular, it can be used to support targeted preventive work with specific schools in order to reduce the risk of exam rule violations.

The technique of ensuring of interoperability for Grid-systems and systems of cloud computing is provided. The technique is constructed on a basis of the uniform approach of ensuring interoperability for systems of the wide class offered by authors and recorded in the national Russian Federation standard.