The article addresses the implementation of Multi-client AI architectures in corporate environments. Against the backdrop of the rapid proliferation of artificial intelligence across organizational ecosystems, the discussed framework gains particular significance-primarily as a key enabler for scaling intelligent solutions without duplicating infrastructure. The relevance of the topic is driven by the growing demand for individualized AI services built upon a unified technological foundation, as well as the imperative for strict data isolation when serving multiple clients-both within a single enterprise and in B2B models. The aim of the study is to identify the structural characteristics of the multi-client approach in AI, its technological and organizational underpinnings, and the key challenges that hinder its sustainable deployment and operation. The analysis reveals conceptual discrepancies in how contemporary researchers interpret the management of cross-client contexts and evaluate multi-tenant AI platforms. Existing literature is predominantly focused either on superficial case studies or on narrow technical details, lacking systematic treatment of managerial dimensions. The novelty of the article lies in the proposed integrative conceptual framework that synthesizes principles of Parameter-Efficient Fine-Tuning (PEFT), Federated Learning, prompt engineering, and RetrievalAugmented Generation (RAG) architectures within the context of enterprise AI adoption. The study contributes to the existing body of knowledge by offering a structured overview of the challenges and opportunities associated with Multi-client AI. The presented material is intended to support professionals in digital transformation, AI platform architects, corporate IT executives, and researchers engaged in designing scalable AI solutions.

In the context of the growing complexity of cyber threats and the increasing frequency of cyberattacks, the relevance of developing effective tools for automatic detection of intrusions into information systems is steadily increasing. Traditional protection methods based on signature analysis often fail to cope with new, unknown attacks, which stimulates the study of alternative approaches, including machine learning (ML) methods. The article discusses the use of ML to solve the problem of binary classification of network activity into normal and abnormal (cyberattack) using the KNIME software platform, which is focused on the visual design of analytical pipelines. The study was conducted on the publicly available Cybersecurity Intrusion Detection dataset, which includes 11 features characterizing network traffic and user behavior (activity time, data transfer protocol, packet size, number of connections, request frequency, etc.). The aim of the work was a comparative study of the efficiency of five ML algorithms: Decision Trees, Naive Bayes, Random Forest, Gradient Boosted Trees, and Simple Regression Tree. The choice of models is due to their prevalence in anomaly detection problems and differences in operating principles: from simple decision rules (Decision Trees) to ensemble methods (Random Forest, Gradient Boosted Trees) that combine several "weak" models to improve accuracy. To prepare the data for training, the principal component analysis (PCA) was used, which allowed us to reduce the dimensionality of the feature space from 11 to 3 components without significant loss of information. This is an important step, since redundancy or correlation of features can negatively affect the quality of the models. Tuning hyperparameters (e.g. tree depth, number of trees in Random Forest, learning rate in Gradient Boosted Trees) and combating overfitting were performed through cross-validation, which ensured the stability of the results on new data. Experiments showed that the highest accuracy (83.055%) and area under the ROC curve (AUC = 0.811) were demonstrated by the Random Forest and Gradient Boosted Trees algorithms. These results are explained by the ability of ensemble methods to take into account nonlinear dependencies in data and their resistance to noise, which is critical for cybersecurity tasks, where attacks can be disguised as normal activity. Decision Trees and Simple Regression Tree models showed lower metrics (accuracy ~75-78%), which is due to their tendency to overfitting on small datasets. Naive Bayes, which assumes independence of features, also gave way to ensemble methods, which confirms the limitations of the independence assumption in the context of network data. Particular attention is paid to the stages of work: from loading and visualizing data (analysis of class distribution, correlations between features) to training models and interpreting the results. The use of KNIME simplified the implementation of the pipeline: the platform provides visual tools for data preprocessing, model tuning and assessing their quality, which makes the approach accessible to cybersecurity specialists without deep programming knowledge. The results of the study contribute to the development of practical applications of ML for protecting information systems. They demonstrate that ensemble methods such as Random Forest and Gradient Boosted Trees can be effectively used to detect cyberattacks in real time, especially in conditions of a limited data set. Promising areas for further work include expanding the dataset to include new types of attacks (e.g. Advanced Persistent Threats, IoT attacks, attacks on cryptographic protocols), as well as integrating deep learning methods (e.g. recurrent neural networks for analyzing network event sequences) to improve the accuracy and adaptability of intrusion detection systems.

The article discusses the use of neural networks to automate user interface (UI) testing. Neural networks, due to their ability to learn and adapt, significantly increase the efficiency and accuracy of testing processes. The paper highlights various approaches to the application of neural network technologies in the automation of UI testing, including the generation and optimization of test scenarios, classification and analysis of defects, as well as forecasting test results. Special attention is paid to the advantages that neural networks provide in improving user interface testing processes, reducing development time and increasing the quality of the final product. However, despite the obvious benefits, the introduction of these technologies is associated with a number of challenges that require additional research and resources. The article emphasizes the importance of integrating neural networks into modern development processes to achieve high software quality standards.

This paper examines methods for detecting malicious applications and software for the Android operating system using machine learning. Provides an analysis of a dataset containing binary features, such as permissions and API method calls that classify applications as malicious or safe. Various machine learning models are reviewed, including Random Forest, SVM, Logistic Regression, and Gradient Boasted Trees, with an assessment of their accuracy and effectiveness. The results demonstrate that ensemble methods and hybrid approaches achieve high accuracy (up to 98%), which confirms their applicability to solve the problem of detecting malware. A special contribution of the work is a detailed analysis of the significance of the features: the most informative for classification were permissions related to access to confidential data (for example, READ_SMS, ACCESS_FINE_LOCATION) and calls to methods that perform potentially dangerous actions (for example, exec). At the same time, hybrid models combining static resolution analysis and dynamic behavior analysis showed the best results. The study also highlights the relevance of developing modern methods to protect mobile devices and suggests avenues for further research, such as a combination of static and dynamic analysis, as well as the application of deep learning to identify complex patterns of malicious behavior. The findings could be used to build better cybersecurity systems capable of adapting to ever-evolving threats. Experiments revealed that traditional signature analysis methods are inferior to machine learning in the effectiveness of detecting new, previously unknown threats. At the same time, an important advantage of the algorithms used was their ability to identify complex relationships between features that are not obvious during manual analysis. For example, combinations of certain permissions with specific API calls can serve as reliable indicators of malicious behavior, even if individually these signs seem harmless. Particular attention is paid to the problem of interpretability of models. Despite the high accuracy of complex algorithms like Gradient Boasted Trees, their black box nature can make it difficult to understand the reasons for classification. To solve this problem, an approach has been proposed to visualize the importance of features and analyze specific cases of misclassification. This allows not only to improve models, but also to provide understandable explanations for security professionals.

With the rapid growth of network traffic and the complexity of the cyber threat structure, traditional methods of analysis and monitoring are becoming less effective, which creates serious vulnerabilities in information security systems. Existing signature-based approaches are unable to cope with new types of attacks, especially those that use polymorphic malware, exploit zero-day vulnerabilities or use stealthy penetration methods. This necessitates the implementation of more intelligent and adaptive solutions that can effectively analyze large volumes of heterogeneous data in real time and promptly detect anomalies in network traffic. The presented study is aimed at developing and testing modern methods for detecting anomalies in network traffic based on machine learning algorithms, including both classical and advanced approaches - deep neural networks and ensemble methods. The goal of the project is to create effective models that can automatically identify suspicious activity, classify possible threats and predict potential attack vectors. Particular attention is paid to the practical application of these models in real time without a significant increase in computational costs. To implement the tasks set, the work uses the KNIME (Konstanz Information Miner) analytical platform, which provides a visual environment for building, training and testing machine learning models. Thanks to its modular architecture and intuitive interface, KNIME simplifies the process of data processing, building workflows and interpreting results. This makes the platform especially attractive to specialists who do not have deep knowledge of programming, but are interested in using intelligent data analysis to improve cybersecurity.

The article examines the role of information technology (IT) as a tool for the social inclusion of people with disabilities. The key areas of application of OT in ensuring access to education, employment, communication, daily life and public services are analyzed. Special attention is paid to assistive technologies, remote forms of interaction and the development of digital infrastructure that promotes the integration of people with disabilities into society. Modern challenges of digital inclusion are highlighted, such as the need for universal design, digital literacy, and compliance with accessibility standards. The article presents the author's scheme reflecting the ecosystem approach to the use of OT in the interests of inclusive development. The conclusion is made about the strategic importance of digital solutions for the formation of an inclusive and fair society.

TDistributing datasets containing DDoS attack patterns in software-defined networks (SDNs) is an important task for improving threat detection methods. This paper examines machine learning (MO) approaches for generating synthetic data based on the HLD-DDoSDN dataset. Generative adversarial networks (GAN), variational autoencoders (VAE), resampling methods (SMOTE, ADASYN), and diffusion models are considered. A comparative analysis of methods is carried out in terms of generation quality, computational complexity, and applicability in DDoS attack detection tasks. The results show that hybrid methods combining GAN and SMOTE demonstrate the best balance between data quality and efficiency.

The article is devoted to the urgent problem of ensuring accurate and stable temperature conditions in conveyor-type tunnel furnaces, widely used in various industries for continuous heat treatment of materials. The efficiency and quality of processing directly depend on maintaining the temperature in each zone of the furnace within the specified limits. Deviations from the set parameters lead to a deterioration in the quality of the final product, an increase in the percentage of defects and inefficient consumption of energy resources.  In this regard, the development and implementation of effective automatic temperature control systems is an important and urgent task to increase productivity and reduce production costs. Existing automatic temperature control systems are often based on classical proportional-integral-differential (PID) controllers with fixed parameters. However, such systems do not always provide optimal control in the dynamically changing conditions of the production process and under the influence of external disturbances, such as fluctuations in mains voltage, changes in conveyor speed or fluctuations in ambient temperature.  This leads to a decrease in the efficiency of the furnace and instability of the temperature regime. The presented article proposes an approach to the development of an automatic temperature control system for a conveyor-type tunnel furnace based on the use of PID controllers with adaptive parameter setting. The adjustment of the controller parameters allows the system to automatically adjust to the changing operating conditions of the furnace and compensate for the influence of external disturbances, providing more stable and accurate temperature control.  The proposed approach is designed to solve the problem of inefficiency of traditional PID controllers with fixed parameters.

The Socialist Republic of Vietnam is a key agricultural country where many valuable export crops are grown. However, their commercial value is limited due to insufficient attention to the conditions of storage of raw materials and their processing, therefore, the application and investment in advanced methods of its preservation and processing is a very relevant area of modern scientific research. Recent research conducted by Vietnamese scientists in collaboration with the international scientific community has shown that applying various innovative technologies to agricultural raw materials, combined with the use of biodegradable polymer packaging materials during storage, leads to a significant reduction in losses during transportation, as well as the development of final products, meeting the demand for food throughout the country. and by increasing their export volumes. However, for the rational implementation of technologies for the production of certain end products from agricultural raw materials or its long-term storage, information on its physico-chemical properties is necessary, and if we consider that most technologies are associated with thermal effects on raw materials, ignorance of thermophysical characteristics becomes an important industry problem, which served to set the goal of this study, which is to determining the thermophysical characteristics of one of the most controversial fruits grown in Vietnam – durian and promising alginate-based packaging material. Within the framework of the conducted research, the specific heat capacity, thermal conductivity coefficients and thermal conductivity of the objects of study were determined, taking into account their humidity, using well-known and adapted techniques. In this study, all the obtained results of determining the thermophysical characteristics for the studied objects are also presented in the form of corresponding graphical and mathematical dependencies.

The article analyses the quality indicators of composite mixes, doughs and bakery products based on first-grade wheat flour and beetroot powder in doses of 3-11% of the flour weight. Straight and low-temperature fermentation methods were used to obtain the bakery products. With an increase for additive, an increase in the titratable acidity of the composite mixtures was observed, the amount of raw gluten decreased while its elasticity was improved. The enriching powder did not have a significant effect on the physic-chemical characteristics of the dough and finished products obtained by the straight method. The quality indicators of the samples prepared under cold fermentation conditions at a temperature of +5 °C and a duration of 20 hours were slightly lower. It was observed that after 18 hours of the experiment, the increase in the acidity of the dough was slower and by the 20th hour it was completely fermented. These data are consistent with the results obtained in the course of a model experiment to study the dynamics of fermentation. The product that was chosen as the best according to the tasting results was the one that included 7 % of table beet powder from the mass of wheat flour. By reducing the sugar dosage in the recipe of this sample by 50 % and resting the dough before fermentation, its consumer properties were improved.  

Introduction. The article presents existing technologies for food films.

Materials and methods. The study is based on content analysis of scientific literature by searching  https://elibrary.ru/, https://www.sci-hub.ru/, https://cyberleninka.ru/, https://scholar.google.com, https://pubmed.ncbi.nlm.nih.gov/. 

Results and discussion. In the course of the work, it was found that edible packaging films are made from carbohydrates (starch, cellulose, etc.), proteins (collagen, gelatin, etc.) and fats, divided into digestible and indigestible. Hydrocolloids are used as a base, and antioxidants, antimicrobial substances, etc. are used as additives. Molding takes place by continuous or intermittent method, as well as by extrusion. Characteristics such as homogeneity, viscosity, etc. are monitored. The films are used for packaging various products, increasing their shelf life due to their antibacterial and antioxidant properties, often using plant extracts.

Conclusion. Based on the results of the study, it can be concluded that Biodegradable food films are environmentally friendly, but less durable than synthetic analogues. Films with antimicrobial properties prolong the shelf life of products, but require careful development. Modern technologies make it possible to create films with improved characteristics (for example, gas permeability). The choice of film depends on the product, storage conditions and environmental requirements; future research is aimed at creating more efficient and environmentally friendly solutions.

Introduction. A by-product of the brewing industry, brewer's grain, which has significant nutritional value, is a potential source for the creation of functional foods. The article provides an analytical review of the current state of research on the biologically active compounds of brewer's grain, the features of various extraction methods, the qualitative characteristics of the extracted components, as well as issues of possible practical application of these substances.

Materials and methods.The study was carried out by searching for specialized scientific literature by keywords in the Scopus, PubMed, Web of Science, Google Scholar databases and on the Elibrary.ru platform; these time frames covered publications of the results of 2020-2025. The work used the methods of scientific search and first screening of publications, data collection and analysis, as well as their systematization and generalization. A total of 44 scientific publications on case studies were analyzed as a data source.

Results and discussion.The results of this study showed that brewer's grain contains a wide range of bioactive compounds with the potential for use as sustainable food and nutraceutical components. The biological activity of compounds extracted from brewer's grains largely depends on the extraction method used. It has been established that brewer's grain protein has a regulatory role in the control of arterial pressure, promoting the inhibition of angiotensin-converting enzyme (ACE) and reducing blood lipid levels. In addition, amino acid residues in the composition of ACE proteins determine their antioxidant and anti-inflammatory properties, and brewer's grain polyphenols have a positive effect on glucose absorption. Brewer's grains contain many minerals, such as Ca, Fe and B1 vitamins with a high degree of bioavailability, which is confirmed by simulation studies of the digestion process.

Conclusion.Along with the growing interest in the use of brewer's grain extracts in the food industry, there is an urgent need for additional studies to comprehensively assess their toxic and allergenic effects. Further development and improvement of extraction technologies is also needed to improve the efficiency of compound isolation and ensure the required level of product safety when included in food products.

Introduction. The presented material highlights the analysis of trends in the development of intellectual property in the field of biotechnology, including food technologies and bioeconomy. Key trends in patent activity and dynamics of intellectual property objects have been identified, which will help to understand the effectiveness of the university's scientific activities and its contribution to the development of the bioeconomy, as well as address the problem of assessing the effectiveness of the university's scientific activities through intellectual property. The author obtained data from 2020 to 2024 on all intellectual property objects in the fields of biotechnology and food technology at Far Eastern Federal University, a leading university in the Eurasian region. The following conclusions were drawn: despite a decline over the five-year period in the registration of new intellectual property objects, patents are primarily based on intellectual property in the areas of food technology and biotechnology, with an increase also seen in certificates for computer programs and databases in the field of bioeconomy. It was concluded that there is indeed a trend comparable to the national one towards increasing the number of intellectual property objects in the fields of biotechnology and bioeconomy within the university, with food sciences leading in terms of registered patents, possibly due to both the established culture of paying special attention to patenting and the development of food technologies and biotechnology within the university and the country as a whole. The article will be useful to researchers dealing with issues of innovation and commercialization of scientific achievements, as well as specialists in the fields of biotechnology and food industry.

Materials and methods. The study is based on comparative analysis, database analysis, and interpretation of the results.

Results and discussion. The research found a slight decrease in the share of registered patents in 2022-2023; however, biotechnology and food technologies have remained leaders in patenting over the past four years, highlighting their importance for maintaining the scientific effectiveness of the university as well as for the agro-industrial sector and food security. Despite the overall decline in patents, the university is actively developing biotechnology projects, aligning with global trends and the significance of food systems economics for the region.

Conclusion. Based on the conducted research, it can be concluded that according to the analysis, intellectual property objects in the field of agro-food biotechnology and food technology form the core of intellectual property at DVFU, which corresponds to the global and nationwide trend of intensive development in food sciences and agro-food biotechnologies. 

The relevance of the use of milk thistle and peppermint extracts in the technology of functional products is substantiated. The effect of plant extracts based on whey on the growth of probiotic microorganisms (Bifidobacterium longum, Propionibacterium freudereichii) was studied. Experimental data on the dynamics of lactic acid synthesis during skim milk fermentation by cultures of probiotic microorganisms have been obtained.

The paper considers soybean and sunflower meal, as well as rapeseed cake as potential sources of complete protein of plant origin for obtaining food protein supplements. A comparative characteristic of the physicochemical and biochemical parameters of the specified types of secondary raw materials of oil production is carried out. Application of the capillary electrophoresis method to identify the amino acid profile. It has been established that soybean meal has the highest concentration of amino acids, including lysine, methionine, isoleucine and glutamic acid, which indicates its high biological value. Sunflower meal occupies an intermediate position, while rapeseed cake is characterized by a reduced content of some amino acids and requires additional processing due to the presence of antinutritional substances. Technological approaches to obtaining protein concentrates using extraction, enzymatic hydrolysis and membrane filtration are presented. The data obtained confirm the prospects of using the studied raw materials as a functional component in food production and are aimed at implementing the principles of resource conservation and sustainable development of the agro-industrial complex.

As part of the study, an analysis of potentially hazardous factors in the production of fermented milk products enriched with a selenium-containing nano-sized system was carried out. All technological stages of production were analyzed and, using an algorithm for selecting critical control points, all categories of potential risk were identified in accordance with the principles of the HACCP system. As a result, a list of hazardous factors was formed and critical control points were established during the production of fermented milk products.

This article presents an analysis of research that reveals the importance and role of electoral culture, the connection between political culture and electoral culture, reflecting the preferences and interests of the electorate; three levels of electoral culture that make up the electoral process at the present stage of development of Russian society are identified. Special attention in the study of electoral culture was paid not only to the subjects of the electoral/electoral process, but also to the value principles that influence the change of socio-political processes in the political structure of the state.

The problem of optimizing the structure and increasing the efficiency of fleet operation in the Russian Federation remains relevant, despite the high level of mechanization of earthworks. This issue is especially acute in conditions of dense urban development, where increased demands are placed on the accuracy and speed of excavation work. In the course of the study of best practices in the field of earthworks, an analysis of existing methods of soil processing, their features and application possibilities was carried out. Based on the data obtained, a comprehensive methodology was developed aimed at ensuring optimal compliance between earthmoving, earthmoving and auxiliary equipment, as well as technological processes. A method for selecting a rational set of machines for excavation in urban conditions is proposed, based on identifying the positions and dependencies that ensure a rational correspondence between the parameters of technological processes. A method for selecting a rational set of machines for excavation in urban conditions is proposed, based on identifying the positions and dependencies that ensure a rational correspondence between the parameters of technological processes and the design of machines. This technique is a key element in the formation of rational sets of machines adapted to the specifics of specific types of earthworks. Special attention was paid to the selection and systematization of leading machines for performing various types of excavation work. This approach makes it possible to maximize the efficiency of using machinery, minimize costs and improve the quality of excavation work, which is especially important in conditions of dense urban development.

Visual content is a powerful tool for manipulating public opinion, especially in election campaigns. Modern technologies, including artificial intelligence (AI), are actively used to create visual images that can influence the perception of the audience. Thanks to the ability of visual content to be remembered and evoke emotional reactions, political technologists can purposefully shape the image of a candidate by manipulating public opinion in their favor. AI allows you not only to accelerate the creation of visual material, but also to analyze the preferences of the audience, adapting the content to specific social groups. However, the use of AI in politics raises ethical questions related to the possibility of manipulation and distortion of reality. Visual content created using AI requires strict controls to ensure transparency and integrity of election processes.

The article examines the consistent development of relations between the Russian Federation and Qatar. After a rather difficult path to mutual understanding, since 2013, cooperation between the countries has reached a fundamentally new level and has served as a basis for bilateral cooperation. Russia's military and political successes in Syria, as well as Moscow's support for Qatar during the diplomatic crisis, have served as starting points for expanding cooperation in the fields of investment, education, healthcare, sports, and military-technical cooperation.

The article is devoted to the analysis of the emergence and implementation of such a social phenomenon as cancel culture. Particular attention is paid to the practice of cancellation by the Baltic countries in relation to Russia, where this phenomenon acts as an instrument of political and cultural confrontation. The author analyzes the process in which the desire to protect rights and freedoms develops into open aggression and leads to further discussions. It is concluded that cancel culture can serve not only as a means of bringing to justice, but also contributes to discrimination, which calls into question the possibility of a constructive dialogue between the conflicting parties. An analytical study was conducted to study the sanctions of Latvia, Lithuania and Estonia against the Russian Federation. The result of the analysis is an understanding of the complex balance between the protection of values and the risks associated with escalating the conflict.

The article presents the main approaches to the study of the phenomenon of electronic voting: legal, sociological, political science and technological. The definitions of electronic voting are presented, and its key parameters are outlined. The evolution of the institute of electronic voting in Russia and in the world is considered. The current status of the institution of electronic voting in the country is described. The problematic aspects and barriers to the implementation of the concept of electronic voting in Russia are described: the "digital divide", the lack of high-speed Internet connection throughout the country, the low level of digital competencies among a certain proportion of the Russian population, the risks for poor citizens and the elderly, the low level of public confidence, the inability to organize public control over voting, the impossibility of ensuring the secrecy of voting, the technological and technical complexity in organizing remote voting, and the vulnerability of the system to external influences.

Youth policy is strategically oriented towards a new vector of development. The study of promising areas of youth self-realization in the region is of particular interest for research. For this purpose, a model of instrumentalization and technological implementation of regional youth policy has been developed. Structurally, the model is represented by theoretical, methodological and technological blocks. The theoretical and methodological components include the following: purpose, objectives, approaches, methods, factors, criteria, conditions, mechanisms. The technology block includes technologies for implementing regional youth policy. The study analyzes the following strategies: a value strategy, a strategy for developing socio-economic potential, and an information strategy. The youth policy of the Lipetsk region has become the subject of research in strategic and technological directions. A value strategy, a strategy for developing socio-economic potential, and an information strategy were considered. In accordance with the strategies, technologies for the implementation of state youth policy are determined: value-based, technology for the formation of a young family, technology for health conservation, technology for the development of socio-economic potential, information technology. The results of the study of strategies and technologies for the regional development of youth policy are interesting for students studying sociological and political disciplines, as well as for all interested persons engaged in research in the field of political processes. The results of the study can be used for lectures in higher education institutions and serve as a guide for further research in this area and the basis for the development of programs and projects in the field of youth policy.

The article examines the problem of continuity and traditions in Russian foreign policy, as well as the process of formation of a new Russian diplomacy in the post-Soviet space in 1991-2000. The author concludes that the new Russian diplomacy has inherited the diplomatic traditions of the Russian Empire and the Soviet Union, while successfully forming its own distinctive features. By the end of the 1990s, Russia's foreign policy towards the post-Soviet countries had become more pragmatic, which made it possible in the following decades to build a system of mutually beneficial relations both in a bilateral format and within the CIS. Modern Russia is persistently striving to build a democratic system of international relations that meets the real needs of the world community in the 21st century. The fundamental principle of this system should be the solution of international problems on a collective basis in strict accordance with the norms of international law.

The article examines the manifestations of extremist sentiments on the Internet. The problem of defining the typology of political extremism and its manifestations in the Internet environment is being solved. The purpose - to analyze the manifestations of various types of political extremism in the online sphere and describe technological recognition algorithms - was solved using the following tasks: the characteristic of political extremism is presented, the typologies of political extremism are analyzed, the manifestations of various types of extremism in the Internet environment are investigated, technological recognition algorithms are disclosed. The main dangerous possibilities of network technologies are identified, and various currents of extremism that realize their goals in the online space are compared. In accordance with federal law, extremist activity is inciting hatred and discord on political, ideological, racial, national or religious grounds, as well as propaganda of superiority or inferiority on the specified grounds. According to the definition of Federal Law No. 114, the features of extremism are the violation of territorial integrity, constitutional order, obstruction of the exercise of civil rights or the activities of state bodies, as well as calls for this. In this study, we use the typology of the famous researcher M.K. Archakov from the dissertation "Political Extremism in Russia", based on several signs of political extremism. According to the structure, hierarchical and decentralized, network organizations are distinguished. The first type can be represented by the classical form of a political party. The second, for example, by a radical youth subculture. A. A. Vasiliev emphasizes the diversity of various forms of extremism on the network. Among them, we can highlight: involving citizens in extremist networks; shaping public opinion and planning mass riots; personal work to recruit people for terrorist acts; inciting hatred and discord on political, religious, national grounds; dissemination of harmful materials. Information technology is a neutral tool. Thus, neural networks can be used to counter extremism. With their help, it is possible to create hardware algorithms for searching for radical materials and statements. It is necessary to assess the risks of new communication systems in advance, as well as to adapt digital tools for the purposes of maintaining law and order. A method of searching for extremist sentiments was formed, consisting of several stages: formation of the research base: selecting a segment of the Internet and collecting specific pages for analysis; technical and hardware search. Collecting additional information about the page and the user: IP address, links, involvement in large- scale networks; analysis of specific materials. Search for extremist statements and messages in various communicative forms, in images, in audio and video materials; social analysis. Search for people and groups that spread extremist sentiments, as well as monitoring those who perceive it. Preparation of a dossier. The strength of this algorithm is that it can be used by an ordinary law enforcement officer in the "manual" search for extremist sentiments, as well as by a program or artificial intelligence. The research results can be applied in the educational process of higher education institutions. 

The article is devoted to the analysis of the use of innovative tools in urban planning, in particular, integrated territorial development (CDT) and master planning. The study examines in detail the advantages and disadvantages of the latter approach. Master planning is a strategically oriented method of managing the spatial development of urban areas. This approach has a number of significant advantages, which makes it in demand among various stakeholders, including government agencies, the private sector, the local population and the expert community. A key aspect of master planning is its ability to ensure effective communication between all participants in the process. The document created within the framework of this approach integrates conceptual approaches to the spatial development of the city, industry strategies and specific tasks, allowing to form a holistic and integrated vision of future development.

Modern construction strives to integrate cutting-edge technologies and design solutions to ensure not only functionality, but also aesthetic appeal of objects. Frame houses, having a number of advantages in terms of economy and speed of construction, require a careful approach to their exterior decoration. Facade finishing plays a key role in protecting the structure from environmental influences and creating a unique image of the building. In this paper, various technologies and styles of exterior finishing of frame houses will be considered, as well as their impact on the overall perception and performance of buildings. The role of design and architectural elements in the design of facades is difficult to overestimate. Current trends in construction are aimed at using environmentally friendly materials that do not have a negative impact on the environment. Selection of materials for exterior decoration, after the end of the operational period and possible for recycling and recycling. Some materials, such as metal and plastic, can be recycled and reused, which reduces the amount of waste and reduces the burden on the environment. Choosing the right materials and technologies allows you not only to protect the building from environmental impact, but also to create comfortable and attractive housing that will delight its owners for many years.