Online Social Networks and Media最新文献

Relevant socio-psychological processes can be detected in social networks thanks to an analysis of linguistic markers that sheds light on the characteristics and dynamics of the social discourse. Usually, linguistic markers comprise a list of words representative of a given construct; however, this approach does not account for contextual interdependencies of words, which can amplify or diminish the relevance of a particular word. In this paper, we present and leverage a scalable method called PageRank-like marker projection (PLMP) that addresses this problem. Its rationale, inspired by PageRank, is meant to fully exploit the interdependencies in a semantic network to project markers from a social discourse level (tweets) to its semantic elements (words). We show how PLMP is able to associate markers with specific words from their semantic context, which allows for an even richer interpretation of the online sentiment. We demonstrate the effectiveness of PLMP in practice by considering specific instances of social discourse on Twitter for three exemplary calls to collective action.

Tourism is vital for most historical and cultural cities. In the context of Smart Cities, there are numerous data sources in tourism domain that could be analyzed to monitor and forecast a range of different indicators related to touristic locations and attractions. In this paper, we propose a framework which exploits social media and big data to forecast both online reputation and touristic attraction presences. To this end, some techniques have been tested and proposed on the basis of machine learning, deep learning, causality assessment and explainable Artificial Intelligence, so as to provide evidence of the relevant variables for each prediction and estimation. An approach has been introduced to analyze the explainability of the proposed solutions, i.e., a multilingual sentiment analysis tool for social media data based on transformers to compare data sources as Trip Advisor and Twitter. Furthermore, causality analysis has been performed to evaluate the temporal impact of social media posts and other factors with respect to the number of presences. The work has been developed in the context of Herit-Data, a European Commission funded project on the exploitation of big data for tourism management and based on the Snap4City infrastructure and platform. Herit-Data has developed solutions for 6 major European touristic locations. In this paper, some of the solutions developed for Florence, Italy and Pont du Gard, France, are reported.

The recent interest for decentralised systems and decentralisation of the control over users’ data brings a shift in the way identities and their information are managed. Self Sovereign Identity (SSI) has been proposed as the next generation paradigm for decentralised identity management. Research on SSI is getting more and more traction, focusing mainly on the management of users’ identifiers and on providing a standard way to express and verify credentials. Instead, this paper focuses on the understanding of the role of trust in SSI and it provides new insight into the trust relationships existing between the different SSI actors. Indeed, the analysis of such roles and the relationships existing between SSI actors reveals that the current paradigm suffers from trust issues between the verifier and the issuer of a verifiable credential.

In order to cope this problem, the paper proposes a new multi-layer framework that exploits trust relationships defined by the actors of the SSI standards (verifiers and issuers of verifiable credentials). An implementation of the framework through Solidity smart contracts has been proposed and deployed on both private and public blockchain networks in order to assess its capabilities. In addition, a dataset related to the spread of spam reviews has been exploited to test the benefits and performance of the proposed framework, demonstrating that it is able to improve the reliability of the SSI paradigm in real-world scenario.

As online social networks have become dominant platforms for public discourse worldwide, there is growing anecdotal evidence of a concurrent rise in social antagonisms. Yet, while the increase in polarization is evident, the extent to which these digital communication ecosystems are driving this shift remains elusive. A dominant scholarly perspective suggests that digital social media lead to the compartmentalization of information channels, potentially culminating in the emergence of echo chambers. However, a growing body of empirical research suggests that the mechanisms influencing ideological demarcation are more complex than a complete communicative decoupling of user groups. This study introduces two intertwined principles that elucidate the dynamics of digital communication: First, socio-cognitive biases of social group formation enforce internal congruence of ideological communities by demarcation from outsiders. Second, algorithmic personalization of content contributes to ideological network formation by creating social redundancy, wherein the same individuals frequently interact in various roles, such as authors, recipients, or disseminators of messages, leading to a surplus of shared ideological fragments. Leveraging these insights, we pioneer a computational simulation model, integrating machine learning based on behavioral data and established recommendation technologies, to explore the evolution of social network structures in digital exchanges. Utilizing advanced methods in opinion dynamics, our model uniquely captures both the algorithmic delivery and the subsequent dissemination of messages by users. Our findings reveal that in ambiguous debate scenarios, the dual components of our model are essential to accurately capture the emergence of social polarization. Consequently, our model offers a forward-looking perspective on the evolution of network communication, facilitating nuanced comparisons with empirical graph benchmarks.

In this paper, we investigate the Least Cost Directed Perfect Awareness Problem (LDPAP), a combinatorial optimization problem that deals with the spread of information on social networks. The objective of LDPAP is to minimize the cost of recruiting individuals capable of starting a propagation of a given news so that it reaches everyone. By showing that LDPAP can be regarded as a generalization of the Perfect Awareness Problem, we establish that LDPAP is NP-hard and we then prove that it remains NP-hard even when restricted to directed acyclic graphs. Our contributions also include two integer programming formulations, a heuristic based on the metaheuristic GRASP and a useful lower bound for the objective function. Lastly, we present extensive experiments comparing the efficiency and efficacy of our heuristic and mathematical models both on synthetic and on real-world datasets.

Stress and depression are prevalent nowadays across people of all ages due to the quick paces of life. People use social media to express their feelings. Thus, social media constitute a valuable form of information for the early recognition of stress and depression. Although many research works have been introduced targeting the early recognition of stress and depression, there are still limitations. There have been proposed multi-task learning settings, which use depression and emotion (or figurative language) as the primary and auxiliary tasks respectively. However, although stress is inextricably linked with depression, researchers face these two tasks as two separate tasks. To address these limitations, we present the first study, which exploits two different datasets collected under different conditions, and introduce two multitask learning frameworks, which use depression and stress as the main and auxiliary tasks respectively. Specifically, we use a depression dataset and a stressful dataset including stressful posts from ten subreddits of five domains. In terms of the first approach, each post passes through a shared BERT layer, which is updated by both tasks. Next, two separate BERT encoder layers are exploited, which are updated by each task separately. Regarding the second approach, it consists of shared and task-specific layers weighted by attention fusion networks. We conduct a series of experiments and compare our approaches with existing research initiatives, single-task learning, and transfer learning. Experiments show multiple advantages of our approaches over state-of-the-art ones.

Previous research has highlighted the importance of network structures in information diffusion on social media. In this study, we explore the role of an individual’s social network structure in predicting publicly announced intention of protest participation. Using the case of ecological protests in Russia and applying machine learning to publicly-available VKontakte data, we classify users into protesters and non-protesters. We have found that personal social networks have a high predictive power allowing user classification with an accuracy of 81%. Meanwhile, using all public VKontakte data, including memberships in activist groups and friendship ties to protesters, we were able to classify users into protesters and non-protesters with a higher accuracy of 96%. Our study contributes to the political-participation literature by demonstrating the importance of personal social networks in predicting protest participation. Our results suggest that in some cases, the likelihood of participating in protests can be significantly influenced by elements of a personal-network structure, inter alia, network density and size. Further explanatory research should be done to explore the mechanisms underlying these relationships.

We examine an unexpected but significant source of positive public health messaging during the COVID-19 pandemic—K-pop fandoms. Leveraging more than 7 million tweets related to mask-wearing and K-pop between March 2020 and December 2021, we analyzed the online spread of the hashtag #WearAMask and vaccine-related tweets amid anti-mask sentiments and public health misinformation. Analyses reveal the South Korean boyband BTS as one of the most significant driver of health discourse. Tweets from health agencies and prominent figures that mentioned K-pop generate 111 times more online responses compared to tweets that did not. These tweets also elicited strong responses from South America, Southeast Asia, and interior States—areas often neglected by mainstream social media campaigns. Network and temporal analysis show increased use from right-leaning elites over time. Mechanistically, strong-levels of parasocial engagement and connectedness allow sustained activism in the community. Our results suggest that public health institutions may leverage pre-existing audience markets to synergistically diffuse and target under-served communities both domestically and globally, especially during health crises.

Bots, simply defined as accounts controlled by automation, can be used as a weapon for online manipulation and pose a threat to the health of platforms. Researchers have studied online platforms to detect, estimate, and characterize bot accounts. Concerns about the prevalence of bots were raised following Elon Musk’s bid to acquire Twitter. In this work, we want to stress that crucial questions need to be answered in order to make a proper estimation and compare different methodologies and definitions based on behaviors and activities; otherwise the real questions concerning the health of online platforms will be confounded by disagreements about definitions and models. We argue how assumptions on bot-likely behavior, the detection approach, and the population inspected can affect the estimation of the percentage of bots on Twitter. Finally, we emphasize the responsibility of platforms to be vigilant, transparent, and unbiased in dealing with threats that may affect their users.