Adress: 123 Naberezhnaya Kanala Griboedova
Topic modeling is a popular technique for clustering large collections of text documents. A variety of different types of regularization is implemented in topic modeling. In this paper, we propose a novel approach for analyzing the influence of different regularization types on results of topic modeling. Based on Renyi entropy, this approach is inspired by the concepts from statistical physics, where an inferred topical structure of a collection can be considered an information statistical system residing in a non-equilibrium state. By testing our approach on four models—Probabilistic Latent Semantic Analysis (pLSA), Additive Regularization of Topic Models (BigARTM), Latent Dirichlet Allocation (LDA) with Gibbs sampling, LDA with variational inference (VLDA)—we, first of all, show that the minimum of Renyi entropy coincides with the “true” number of topics, as determined in two labelled collections. Simultaneously, we find that Hierarchical Dirichlet Process (HDP) model as a well-known approach for topic number optimization fails to detect such optimum. Next, we demonstrate that large values of the regularization coefficient in BigARTM significantly shift the minimum of entropy from the topic number optimum, which effect is not observed for hyper-parameters in LDA with Gibbs sampling. We conclude that regularization may introduce unpredictable distortions into topic models that need further research.
The article examines the recent ‘schism’ in Eastern Orthodoxy to show how religion and politics are strongly intertwined in disputes over territory and sovereignty. It argues that two logics are at play in this conflict: one grounded in the theological‐political concept of ‘canonical territory’, the other in the notion of ‘communion’ at the basis of the Christian fellowship. The first is deployed in claims for national sovereignty as well as imperial domination, while the latter can make or break communities of faith. Drawing a parallel between the post‐socialist revival of religion in Ukraine and the current mobilization on the ground, it shows how these contradictory logics shape the fate of people, churches and states.
In this paper, drawing on empirical evidence from Russian society, the authors seek to analyze the ways in which ordinary people can overcome a perceived gap between “high” politics and their everyday experiences. We argue that the concept of everyday politics is not enough to prepare the ground for politicization in everyday life, at least in highly dismantled society. In examining ethnographic case studies of people who consider themselves apolitical, the paper introduces the concept of “pragmatic politics,” which is defined as an activity of inscribing the broader world within the sensitive everyday experience. The examined case studies speak to four different modes of everyday politics that reveal various modalities of pragmatic politics.
This chapter underlines the key role of a city centre in urban space gastronomy. It offers a four-step perspective, ranging from urban to local. First, the example of Saint-Petersburg (Russia) shows that gastronomy reflects the major phases of urban growth. Here, eating establishments are used as a proxy for the city centre. Second, the example of Warsow's Śródmieście district in Poland indicates the constant growth in catering services in this central borough since 1994. Using density analysis, it shows gastronomy hotspots in the centre of the city. Next, the case of Kraków (Poland) focuses on the centre of a historical tourist city, where there has been both quantitative growth in the number of eating establishments and a change in their distribution. The last examples offer a local perspective, specifically they concern the district Żoliborz in Warsaw, Poland, and the neighbourhood of Podskalí in Prague, the Czech Republic, which are near the city centre.
This study analyzes roll call voting in the Council of Ministers from December 2003 to May 2019 in order to identify the factors that determine the strategies of coalition behavior of 28 EU Member States. The analysis makes possible to single out two important cleavages affecting the coalitional preferences of the Ministers of states. The first cleavage is observed between the EU members from Eastern and Western Europe. The second cleavage is associated with the duration of the countries’ EU membership. The rationalistic intentions of member countries related to the agenda of the Council and their ideological preferences also influence the process of coalition formation and allow the EU states to go beyond the geographic and ‘temporal’ cleavages.
The use of social network sites helps people to make and maintain social ties accumulating social capital, which is increasingly important for individual success. There is a wide variation in the amount and structure of online ties, and to some extent this variation is contingent on specific online user behaviors which are to date under-researched. In this work, we examine an entire city-bounded friendship network (N = 194,601) extracted from VK social network site to explore how specific online user behaviors are related to structural social capital in a network of geographically proximate ties. Social network analysis was used to evaluate individual social capital as a network asset, and multiple regression analysis–to determine and estimate the effects of online user behaviors on social capital. The analysis reveals that the graph is both clustered and highly centralized which suggests the presence of a hierarchical structure: a set of sub-communities united by city-level hubs. Against this background, membership in more online groups is positively associated with user’s brokerage in the location-bounded network. Additionally, the share of local friends, the number of received likes and the duration of SNS use are associated with social capital indicators. This contributes to the literature on the formation of online social capital, examined at the level of a large and geographically localized population.
This article addresses the puzzle of electoral engineering in autocracies using data from three rounds of Russian regional legislative elections between 2003 and 2017. The analysis shows that electoral engineering was widespread in regions where governors lacked the resources necessary to rely on blatant forms of electoral malpractice for the benefit of United Russia. This pattern became evident during the third round of regional legislative elections. The study indicates that the manipulation of electoral systems may be important for authoritarian rulers when they are unable to rely on blatant electoral malpractice to ensure the certainty of electoral outcomes.
THe article discusses different perspectives of work-related learning and career trajectories expectations in enhancing the life chances of youth.
Field experiments have provided ample evidence of ethnic and racial discrimination in the labour market. Less is known about how discrimination varies in multi-ethnic societies, where the ethnic composition of populations is different across locations. Inter-group contact and institutional arrangements for ethnic minorities can mitigate the sense of group threat and reduce discrimination. To provide empirical evidence of this, we conduct a field experiment of ethnic discrimination in Russia with a sample of over 9,000 job applications. We compare ethnically homogeneous cities and cities with ethnically mixed populations and privileged institutional status of ethnic minorities. We find strong discrimination against visible minorities in the former but much weaker discrimination in the latter. These findings demonstrate how institutions and historical contexts of inter-group relations can affect ethnic prejudice and discrimination.
In practice, the critical step in building machine learning models of big data (BD) is costly in terms of time and the computing resources procedure of parameter tuning with a grid search. Due to the size, BD are comparable to mesoscopic physical systems. Hence, methods of statistical physics could be applied to BD. The paper shows that topic modeling demonstrates self-similar behavior under the condition of a varying number of clusters. Such behavior allows using a renormalization technique. The combination of a renormalization procedure with the Rényi entropy approach allows for fast searching of the optimal number of clusters. In this paper, the renormalization procedure is developed for the Latent Dirichlet Allocation (LDA) model with a variational Expectation-Maximization algorithm. The experiments were conducted on two document collections with a known number of clusters in two languages. The paper presents results for three versions of the renormalization procedure: (1) a renormalization with the random merging of clusters, (2) a renormalization based on minimal values of Kullback–Leibler divergence and (3) a renormalization with merging clusters with minimal values of Rényi entropy. The paper shows that the renormalization procedure allows finding the optimal number of topics 26 times faster than grid search without significant loss of quality.
In this article we analyze the independent Tatar rap scene in two relevant contexts: the globalization of this musical culture, and the post-Soviet nation-building efforts in the Republic of Tatarstan (Russian Federation). Having analyzed 29 in-depth biographical interviews with young rap scene participants and diary entries obtained in the course of a one-month-long participant observation, we conclude that the Tatar rap scene is a special case in the Tatar urban youth culture shaped by the younger generation of Tatar-speaking intelligentsia (humanities graduates and creative professionals) in opposition to both the cultural policy of Russification of the imperial center (Moscow) and the folklorized version of the Soviet Tatar culture.
The paper considers how the productivity and innovations of Russian regions are associated with the heterogeneity of the population by ethno-linguistic affiliation, as well as by country and region of origin. The study contributes to the corpus of papers on economic impact of cultural diversity with the focus on Russia and addresses questions what are the spatial effects of the impact. We find positive associations of the productivity with population heterogeneity by country and region of origin, and the knowledge generation with heterogeneity within the group of immigrants. Our results also contribute to the economic policy by obtaining positive significant spatial effects for the productivity and knowledge creation for Russian regions and by defining main significant factors for these effects.
This study is focused on the problem of the mismatch of competencies of Masters of Public Administration (MPA) graduates in Russia and current Russian public servants. A mixed methods approach was used to analyze quantitative (n = 734) and qualitative data about the real-world competencies of local, regional, and federal government officials in comparison to what MPA graduates get from their education. The comparison of competency models of MPA graduates and government officials indicated that the most of the competencies are useful in public administration practice, but there is still lot to change in the approach of compiling the content of MPA educational standards and educational programs.
This introduction aims to present the general outline of the special issue and to elaborate on the context against which most of the studies were conducted. We discuss the political, economic, social, and historical processes that contribute to shape Russia; this helps to understand local activism and protest in contemporary Russia. Since this context is relevant to all the papers, the readers would benefit from reading this introduction first. The second part of this paper introduces the contribution that the special issue makes to the study of activism and politics, with papers analyzing different aspects and kinds of activism in a variety of circumstances and settings. A central question common to all papers is the problem of politicization which is treated at the intersection between social and political inequalities, the experience of everyday life and political imagination.
This paper focuses on the Tsoi Wall in Moscow, an iconic place on Russia’s music map that appeared in Moscow in 1990 in memory of the cult Soviet rock musician Viktor Tsoi, to develop a framework for studying non-auratic music place—that is, places that are not connected with the biographies of musicians or musical events, but emerge directly from the experiences of visitors and fans. These places are constantly negotiated and only lightly formalized, but are nevertheless enduring. To analyze this type of place, we propose a concept of institutionalization “in becoming.” The case of the Tsoi Wall reveals that light formalization (vague and changing positions and rules, and openness to different interpretations of a place and ways of using it) leads to the recognition of the place as a significant one and to its popularity. We put institutionalization “in becoming” in a wider context and juxtapose it with well-studied musical places in Europe and the US.
The article deals with the ways Russian authorities have constructed the social problem of HIV/AIDS (human immunodeficiency virus/ acquired immune deficiency syndrome) in Russia. The statistical construction of HIV/AIDS includes data indicating the significant rise of HIV prevalence in Russia since 2000. The study focuses on what and how Russian authorities speak about HIV/AIDS, while there are official data on the rapid spread of the virus in the country. The work is based on a discourse analysis of the authorities’ rhetoric about HIV/AIDS. During his first presidential terms, Vladimir Putin constructed HIV/AIDS not as an epidemic in the country, but as a “global problem,” representing Russia as a participant in international efforts to combat AIDS. The president problematized the HIV spread through the rhetoric of endangerment but without its crucial term “epidemic,” while at the same time de-problematized HIV in Russia by the strategy of naturalizing (“this is a problem that all countries face”). The Russian authorities appealed to traditional moral values and spoke about marginal or risk groups, rather than risk practices. After the deterioration of relations with Western countries since 2007, the Russian president excluded HIV/AIDS problem from his public agenda, despite the existence of the data on steep HIV growth in Russia. The Russian president’s traditionalism, de-problematization, and silence concerning HIV/AIDS lead to the absence of the HIV/AIDS issues in media agenda, the agenda of local authorities, and consequently the personal agendas of Russian citizens. The consequences are ignorance, fears, stigmatization of people living with HIV, semi-legal status of needle, and syringe exchange programs for intravenous drug users, low antiretroviral therapy coverage, and the continuing HIV epidemic.
Social science faces tremendous growth of available data about social phenomena on the Internet; however, social science students are usually not prepared to challenges and opportunities of analyzing online data. One of the areas where this growth is especially important is social studies of consumption. In this article we discuss a prototype of the visualization tool intended to support learning netnographic analysis with computational tools
In practice, to build a machine learning model of big data, one needs to tune model parameters. The process of parameter tuning involves extremely time-consuming and computationally expensive grid search. However, the theory of statistical physics provides techniques allowing us to optimize this process. The paper shows that a function of the output of topic modeling demonstrates self-similar behavior under variation of the number of clusters. Such behavior allows using a renormalization technique. A combination of renormalization procedure with the Renyi entropy approach allows for quick searching of the optimal number of topics. In this paper, the renormalization procedure is developed for the probabilistic Latent Semantic Analysis (pLSA), and the Latent Dirichlet Allocation model with variational Expectation–Maximization algorithm (VLDA) and the Latent Dirichlet Allocation model with granulated Gibbs sampling procedure (GLDA). The experiments were conducted on two test datasets with a known number of topics in two different languages and on one unlabeled test dataset with an unknown number of topics. The paper shows that the renormalization procedure allows for finding an approximation of the optimal number of topics at least 30 times faster than the grid search without significant loss of quality.