Serina Chang, Alicja Chaszczewicz, Emma Wang, Maya Josifovska, Emma Pierson, Jure Leskovec
Generating social networks is essential for many applications, such as epidemic modeling and social simulations. Prior approaches either involve deep learning models, which require many observed networks for training, or stylized models, which are limited in their realism and flexibility. In contrast, LLMs offer the potential for zero-shot and flexible network generation. However, two key questions are: (1) are LLM's generated networks realistic, and (2) what are risks of bias, given the importance of demographics in forming social ties? To answer these questions, we develop three prompting methods for network generation and compare the generated networks to real social networks. We find that more realistic networks are generated with "local" methods, where the LLM constructs relations for one persona at a time, compared to "global" methods that construct the entire network at once. We also find that the generated networks match real networks on many characteristics, including density, clustering, community structure, and degree. However, we find that LLMs emphasize political homophily over all other types of homophily and overestimate political homophily relative to real-world measures.
{"title":"LLMs generate structurally realistic social networks but overestimate political homophily","authors":"Serina Chang, Alicja Chaszczewicz, Emma Wang, Maya Josifovska, Emma Pierson, Jure Leskovec","doi":"arxiv-2408.16629","DOIUrl":"https://doi.org/arxiv-2408.16629","url":null,"abstract":"Generating social networks is essential for many applications, such as\u0000epidemic modeling and social simulations. Prior approaches either involve deep\u0000learning models, which require many observed networks for training, or stylized\u0000models, which are limited in their realism and flexibility. In contrast, LLMs\u0000offer the potential for zero-shot and flexible network generation. However, two\u0000key questions are: (1) are LLM's generated networks realistic, and (2) what are\u0000risks of bias, given the importance of demographics in forming social ties? To\u0000answer these questions, we develop three prompting methods for network\u0000generation and compare the generated networks to real social networks. We find\u0000that more realistic networks are generated with \"local\" methods, where the LLM\u0000constructs relations for one persona at a time, compared to \"global\" methods\u0000that construct the entire network at once. We also find that the generated\u0000networks match real networks on many characteristics, including density,\u0000clustering, community structure, and degree. However, we find that LLMs\u0000emphasize political homophily over all other types of homophily and\u0000overestimate political homophily relative to real-world measures.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142227856","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Federated graph learning (FGL) has emerged as a promising distributed training paradigm for graph neural networks across multiple local systems without direct data sharing. This approach is particularly beneficial in privacy-sensitive scenarios and offers a new perspective on addressing scalability challenges in large-scale graph learning. Despite the proliferation of FGL, the diverse motivations from practical applications, spanning various research backgrounds and experimental settings, pose a significant challenge to fair evaluation. To fill this gap, we propose OpenFGL, a unified benchmark designed for the primary FGL scenarios: Graph-FL and Subgraph-FL. Specifically, OpenFGL includes 38 graph datasets from 16 application domains, 8 federated data simulation strategies that emphasize graph properties, and 5 graph-based downstream tasks. Additionally, it offers 18 recently proposed SOTA FGL algorithms through a user-friendly API, enabling a thorough comparison and comprehensive evaluation of their effectiveness, robustness, and efficiency. Empirical results demonstrate the ability of FGL while also revealing its potential limitations, offering valuable insights for future exploration in this thriving field.
{"title":"OpenFGL: A Comprehensive Benchmarks for Federated Graph Learning","authors":"Xunkai Li, Yinlin Zhu, Boyang Pang, Guochen Yan, Yeyu Yan, Zening Li, Zhengyu Wu, Wentao Zhang, Rong-Hua Li, Guoren Wang","doi":"arxiv-2408.16288","DOIUrl":"https://doi.org/arxiv-2408.16288","url":null,"abstract":"Federated graph learning (FGL) has emerged as a promising distributed\u0000training paradigm for graph neural networks across multiple local systems\u0000without direct data sharing. This approach is particularly beneficial in\u0000privacy-sensitive scenarios and offers a new perspective on addressing\u0000scalability challenges in large-scale graph learning. Despite the proliferation\u0000of FGL, the diverse motivations from practical applications, spanning various\u0000research backgrounds and experimental settings, pose a significant challenge to\u0000fair evaluation. To fill this gap, we propose OpenFGL, a unified benchmark\u0000designed for the primary FGL scenarios: Graph-FL and Subgraph-FL. Specifically,\u0000OpenFGL includes 38 graph datasets from 16 application domains, 8 federated\u0000data simulation strategies that emphasize graph properties, and 5 graph-based\u0000downstream tasks. Additionally, it offers 18 recently proposed SOTA FGL\u0000algorithms through a user-friendly API, enabling a thorough comparison and\u0000comprehensive evaluation of their effectiveness, robustness, and efficiency.\u0000Empirical results demonstrate the ability of FGL while also revealing its\u0000potential limitations, offering valuable insights for future exploration in\u0000this thriving field.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This research aims to investigate the effects of information cocooning on group mood changes caused by information spreading. The simulation of the realistic network evolution process is realized at the structural level by building a network evolution model based on individual viewpoints. Abstracting the accuracy of the real intelligent recommendation process by setting RA (Recommended Accuracy). By analyzing the information cocoon effect due to the recommendation in the comment section, we provide the structural basis of spreading for the dynamics model. A dynamics model of emotion spreading is developed to explore the trend of individual emotion spreading and to quantify the change of group emotion. Through experiments and analysis, this paper concludes that the information cocoon has a positive effect on the stability of group emotions, and that the H-CAC (Hidden Comment Area Cocoon) structure exists widely in real online social networks, and can produce a protective "harbor" effect in the competition of public opinion and cognitive games. The validity of the model is verified by comparison with real cases and generalization ability experiments. This work provides a multi-perspective analysis and visualization, providing more quantitative results. The research is expected to provide new perspectives and tools for understanding the reality of information cocooning and expanding the scenarios of its use.
本研究旨在探讨信息茧房对信息传播引起的群体情绪变化的影响。通过建立基于个人观点的网络演化模型，在结构层面上实现了对现实网络演化过程的模拟。通过设置 RA（推荐准确率），抽象出真实智能推荐过程的准确性。通过分析评论区中的推荐所产生的信息茧效应，为动态模型提供了传播的结构基础。本文建立了情感传播动力学模型，以探索个体情感传播的趋势并量化群体情感的变化。通过实验和分析，本文得出结论：信息茧对群体情绪的稳定性有积极作用，H-CAC（Hidden Comment Area Cocoon）结构广泛存在于真实的网络社交网络中，并能在舆论竞争和认知博弈中产生保护性的 "港湾 "效应。该模型的有效性通过与真实案例的对比和泛化能力实验得到了验证。这项工作提供了多视角分析和可视化，提供了更多量化结果。该研究有望为理解信息茧房的现实和拓展其使用场景提供新的视角和工具。
{"title":"IC always bad? : Information Cocooning as a Group Emotional Stabilization Role in Social Networks","authors":"Jinhu Ren, Tianlong Fan, Linyuan Lü, Xifei Fu","doi":"arxiv-2408.16295","DOIUrl":"https://doi.org/arxiv-2408.16295","url":null,"abstract":"This research aims to investigate the effects of information cocooning on\u0000group mood changes caused by information spreading. The simulation of the\u0000realistic network evolution process is realized at the structural level by\u0000building a network evolution model based on individual viewpoints. Abstracting\u0000the accuracy of the real intelligent recommendation process by setting RA\u0000(Recommended Accuracy). By analyzing the information cocoon effect due to the\u0000recommendation in the comment section, we provide the structural basis of\u0000spreading for the dynamics model. A dynamics model of emotion spreading is\u0000developed to explore the trend of individual emotion spreading and to quantify\u0000the change of group emotion. Through experiments and analysis, this paper\u0000concludes that the information cocoon has a positive effect on the stability of\u0000group emotions, and that the H-CAC (Hidden Comment Area Cocoon) structure\u0000exists widely in real online social networks, and can produce a protective\u0000\"harbor\" effect in the competition of public opinion and cognitive games. The\u0000validity of the model is verified by comparison with real cases and\u0000generalization ability experiments. This work provides a multi-perspective\u0000analysis and visualization, providing more quantitative results. The research\u0000is expected to provide new perspectives and tools for understanding the reality\u0000of information cocooning and expanding the scenarios of its use.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214918","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hong Zhou, Peifeng Lai, Zhida Sun, Xiangyuan Chen, Yang Chen, Huisi Wu, Yong Wang
With the increase of graph size, it becomes difficult or even impossible to visualize graph structures clearly within the limited screen space. Consequently, it is crucial to design effective visual representations for large graphs. In this paper, we propose AdaMotif, a novel approach that can capture the essential structure patterns of large graphs and effectively reveal the overall structures via adaptive motif designs. Specifically, our approach involves partitioning a given large graph into multiple subgraphs, then clustering similar subgraphs and extracting similar structural information within each cluster. Subsequently, adaptive motifs representing each cluster are generated and utilized to replace the corresponding subgraphs, leading to a simplified visualization. Our approach aims to preserve as much information as possible from the subgraphs while simplifying the graph efficiently. Notably, our approach successfully visualizes crucial community information within a large graph. We conduct case studies and a user study using real-world graphs to validate the effectiveness of our proposed approach. The results demonstrate the capability of our approach in simplifying graphs while retaining important structural and community information.
{"title":"AdaMotif: Graph Simplification via Adaptive Motif Design","authors":"Hong Zhou, Peifeng Lai, Zhida Sun, Xiangyuan Chen, Yang Chen, Huisi Wu, Yong Wang","doi":"arxiv-2408.16308","DOIUrl":"https://doi.org/arxiv-2408.16308","url":null,"abstract":"With the increase of graph size, it becomes difficult or even impossible to\u0000visualize graph structures clearly within the limited screen space.\u0000Consequently, it is crucial to design effective visual representations for\u0000large graphs. In this paper, we propose AdaMotif, a novel approach that can\u0000capture the essential structure patterns of large graphs and effectively reveal\u0000the overall structures via adaptive motif designs. Specifically, our approach\u0000involves partitioning a given large graph into multiple subgraphs, then\u0000clustering similar subgraphs and extracting similar structural information\u0000within each cluster. Subsequently, adaptive motifs representing each cluster\u0000are generated and utilized to replace the corresponding subgraphs, leading to a\u0000simplified visualization. Our approach aims to preserve as much information as\u0000possible from the subgraphs while simplifying the graph efficiently. Notably,\u0000our approach successfully visualizes crucial community information within a\u0000large graph. We conduct case studies and a user study using real-world graphs\u0000to validate the effectiveness of our proposed approach. The results demonstrate\u0000the capability of our approach in simplifying graphs while retaining important\u0000structural and community information.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Descriptive and inferential social network analysis has become common in public administration studies of network governance and management. A large literature has developed in two broad categories: antecedents of network structure, and network effects and outcomes. A new topic is emerging on network interventions that applies knowledge of network formation and effects to actively intervene in the social context of interaction. Yet, the question remains how might scholars deploy and determine the impact of network interventions. Inferential network analysis has primarily focused on statistical simulations of network distributions to produce probability estimates on parameters of interest in observed networks, e.g. ERGMs. There is less attention to design elements for causal inference in the network context, such as experimental interventions, randomization, control and comparison networks, and spillovers. We advance a number of important questions for network research, examine important inferential challenges and other issues related to inference in networks, and focus on a set of possible network inference models. We categorize models of network inference into (i) observational studies of networks, using descriptive and stochastic methods that lack intervention, randomization, or comparison networks; (ii) simulation studies that leverage computational resources for generating inference; (iii) natural network experiments, with unintentional network-based interventions; (iv) network field experiments, with designed interventions accompanied by comparison networks; and (v) laboratory experiments that design and implement randomization to treatment and control networks. The article offers a guide to network researchers interested in questions, challenges, and models of inference for network analysis in public administration.
{"title":"Network Inference in Public Administration: Questions, Challenges, and Models of Causality","authors":"Travis A. Whetsell, Michael D. Siciliano","doi":"arxiv-2408.16933","DOIUrl":"https://doi.org/arxiv-2408.16933","url":null,"abstract":"Descriptive and inferential social network analysis has become common in\u0000public administration studies of network governance and management. A large\u0000literature has developed in two broad categories: antecedents of network\u0000structure, and network effects and outcomes. A new topic is emerging on network\u0000interventions that applies knowledge of network formation and effects to\u0000actively intervene in the social context of interaction. Yet, the question\u0000remains how might scholars deploy and determine the impact of network\u0000interventions. Inferential network analysis has primarily focused on\u0000statistical simulations of network distributions to produce probability\u0000estimates on parameters of interest in observed networks, e.g. ERGMs. There is\u0000less attention to design elements for causal inference in the network context,\u0000such as experimental interventions, randomization, control and comparison\u0000networks, and spillovers. We advance a number of important questions for\u0000network research, examine important inferential challenges and other issues\u0000related to inference in networks, and focus on a set of possible network\u0000inference models. We categorize models of network inference into (i)\u0000observational studies of networks, using descriptive and stochastic methods\u0000that lack intervention, randomization, or comparison networks; (ii) simulation\u0000studies that leverage computational resources for generating inference; (iii)\u0000natural network experiments, with unintentional network-based interventions;\u0000(iv) network field experiments, with designed interventions accompanied by\u0000comparison networks; and (v) laboratory experiments that design and implement\u0000randomization to treatment and control networks. The article offers a guide to\u0000network researchers interested in questions, challenges, and models of\u0000inference for network analysis in public administration.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142226906","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
International Auxiliary Languages (IALs) are constructed languages designed to facilitate communication among speakers of different native languages while fostering equality, efficiency, and cross-cultural understanding. This study focuses on analyzing the editions of IALs on Wikipedia, including Simple English, Esperanto, Ido, Interlingua, Volapuk, Interlingue, and Novial. We compare them with three natural languages: English, Spanish, and Catalan. Our aim is to establish a basis for the use of IALs in Wikipedia as well as showcase a new methodology for categorizing wikis. We found in total there are 1.3 million articles written in these languages and they gather 15.6 million monthly views. Although this is not a negligible amount of content, in comparison with large natural language projects there is still a big room for improvement. We concluded that IAL editions on Wikipedia are similar to other projects, behaving proportionally to their communities' size. Therefore, the key to their growth is augmenting the amount and quality of the content offered in these languages. To that end, we offer a set of statistics to understand and improve these projects, and we developed a webpage that displays our findings to foster knowledge sharing and facilitate the expansion of the IAL communities.
{"title":"Constructing a Common Ground: Analyzing the quality and usage of International Auxiliary Languages in Wikipedia","authors":"Marta Alet, Diego Saez-Trumper","doi":"arxiv-2408.15873","DOIUrl":"https://doi.org/arxiv-2408.15873","url":null,"abstract":"International Auxiliary Languages (IALs) are constructed languages designed\u0000to facilitate communication among speakers of different native languages while\u0000fostering equality, efficiency, and cross-cultural understanding. This study\u0000focuses on analyzing the editions of IALs on Wikipedia, including Simple\u0000English, Esperanto, Ido, Interlingua, Volapuk, Interlingue, and Novial. We\u0000compare them with three natural languages: English, Spanish, and Catalan. Our\u0000aim is to establish a basis for the use of IALs in Wikipedia as well as\u0000showcase a new methodology for categorizing wikis. We found in total there are\u00001.3 million articles written in these languages and they gather 15.6 million\u0000monthly views. Although this is not a negligible amount of content, in\u0000comparison with large natural language projects there is still a big room for\u0000improvement. We concluded that IAL editions on Wikipedia are similar to other\u0000projects, behaving proportionally to their communities' size. Therefore, the\u0000key to their growth is augmenting the amount and quality of the content offered\u0000in these languages. To that end, we offer a set of statistics to understand and\u0000improve these projects, and we developed a webpage that displays our findings\u0000to foster knowledge sharing and facilitate the expansion of the IAL\u0000communities.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142226905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Identifying (a) systemic barriers to quality healthcare access and (b) key indicators of care efficacy in the United States remains a significant challenge. To improve our understanding of regional disparities in care delivery, we introduce a novel application of curvature, a geometrical-topological property of networks, to Physician Referral Networks. Our initial findings reveal that Forman-Ricci and Ollivier-Ricci curvature measures, which are known for their expressive power in characterizing network structure, offer promising indicators for detecting variations in healthcare efficacy while capturing a range of significant regional demographic features. We also present APPARENT, an open-source tool that leverages Ricci curvature and other network features to examine correlations between regional Physician Referral Networks structure, local census data, healthcare effectiveness, and patient outcomes.
{"title":"Characterizing Physician Referral Networks with Ricci Curvature","authors":"Jeremy Wayland, Russel J. Funk, Bastian Rieck","doi":"arxiv-2408.16022","DOIUrl":"https://doi.org/arxiv-2408.16022","url":null,"abstract":"Identifying (a) systemic barriers to quality healthcare access and (b) key\u0000indicators of care efficacy in the United States remains a significant\u0000challenge. To improve our understanding of regional disparities in care\u0000delivery, we introduce a novel application of curvature, a\u0000geometrical-topological property of networks, to Physician Referral Networks.\u0000Our initial findings reveal that Forman-Ricci and Ollivier-Ricci curvature\u0000measures, which are known for their expressive power in characterizing network\u0000structure, offer promising indicators for detecting variations in healthcare\u0000efficacy while capturing a range of significant regional demographic features.\u0000We also present APPARENT, an open-source tool that leverages Ricci curvature\u0000and other network features to examine correlations between regional Physician\u0000Referral Networks structure, local census data, healthcare effectiveness, and\u0000patient outcomes.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214921","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Media bias significantly shapes public perception by reinforcing stereotypes and exacerbating societal divisions. Prior research has often focused on isolated media bias dimensions such as textit{political bias} or textit{racial bias}, neglecting the complex interrelationships among various bias dimensions across different topic domains. Moreover, we observe that models trained on existing media bias benchmarks fail to generalize effectively on recent social media posts, particularly in certain bias identification tasks. This shortfall primarily arises because these benchmarks do not adequately reflect the rapidly evolving nature of social media content, which is characterized by shifting user behaviors and emerging trends. In response to these limitations, our research introduces a novel dataset collected from YouTube and Reddit over the past five years. Our dataset includes automated annotations for YouTube content across a broad spectrum of bias dimensions, such as gender, racial, and political biases, as well as hate speech, among others. It spans diverse domains including politics, sports, healthcare, education, and entertainment, reflecting the complex interplay of biases across different societal sectors. Through comprehensive statistical analysis, we identify significant differences in bias expression patterns and intra-domain bias correlations across these domains. By utilizing our understanding of the correlations among various bias dimensions, we lay the groundwork for creating advanced systems capable of detecting multiple biases simultaneously. Overall, our dataset advances the field of media bias identification, contributing to the development of tools that promote fairer media consumption. The comprehensive awareness of existing media bias fosters more ethical journalism, promotes cultural sensitivity, and supports a more informed and equitable public discourse.
{"title":"Intertwined Biases Across Social Media Spheres: Unpacking Correlations in Media Bias Dimensions","authors":"Yifan Liu, Yike Li, Dong Wang","doi":"arxiv-2408.15406","DOIUrl":"https://doi.org/arxiv-2408.15406","url":null,"abstract":"Media bias significantly shapes public perception by reinforcing stereotypes\u0000and exacerbating societal divisions. Prior research has often focused on\u0000isolated media bias dimensions such as textit{political bias} or\u0000textit{racial bias}, neglecting the complex interrelationships among various\u0000bias dimensions across different topic domains. Moreover, we observe that\u0000models trained on existing media bias benchmarks fail to generalize effectively\u0000on recent social media posts, particularly in certain bias identification\u0000tasks. This shortfall primarily arises because these benchmarks do not\u0000adequately reflect the rapidly evolving nature of social media content, which\u0000is characterized by shifting user behaviors and emerging trends. In response to\u0000these limitations, our research introduces a novel dataset collected from\u0000YouTube and Reddit over the past five years. Our dataset includes automated\u0000annotations for YouTube content across a broad spectrum of bias dimensions,\u0000such as gender, racial, and political biases, as well as hate speech, among\u0000others. It spans diverse domains including politics, sports, healthcare,\u0000education, and entertainment, reflecting the complex interplay of biases across\u0000different societal sectors. Through comprehensive statistical analysis, we\u0000identify significant differences in bias expression patterns and intra-domain\u0000bias correlations across these domains. By utilizing our understanding of the\u0000correlations among various bias dimensions, we lay the groundwork for creating\u0000advanced systems capable of detecting multiple biases simultaneously. Overall,\u0000our dataset advances the field of media bias identification, contributing to\u0000the development of tools that promote fairer media consumption. The\u0000comprehensive awareness of existing media bias fosters more ethical journalism,\u0000promotes cultural sensitivity, and supports a more informed and equitable\u0000public discourse.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214923","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Conspiracy theories, particularly those focused on anti-vaccine narratives and the promotion of off-label medications such as MMS and CDS, have proliferated on Telegram, including in Brazil, finding fertile ground among communities that share esoteric beliefs and distrust towards scientific institutions. In this context, this study seeks to answer how Brazilian conspiracy theory communities on Telegram are characterized and articulated concerning anti-vaccine themes and off-label medications? It is important to highlight that this study is part of a series of seven studies aimed at understanding and characterizing Brazilian conspiracy theory communities on Telegram. This series of seven studies is openly and originally available on the arXiv of Cornell University, applying a mirrored method across all studies, changing only the thematic object of analysis and providing replicable research, including proprietary and original codes developed, contributing to the culture of free and open-source software. Regarding the main findings of this study, it was observed: Themes such as the New World Order and Apocalypse and Survivalism act as significant gateways to anti-vaccine narratives, connecting them to theories of global control; Globalism and New World Order stand out as the main communities receiving invitations from anti-vaccine communities; Occultism and Esotericism emerge as the largest sources of invitations to off-label medication communities, creating a strong connection between esoteric beliefs and the promotion of non-scientific treatments; Anti-vaccine narratives experienced a 290% increase during the COVID-19 pandemic, evidencing a growing interconnectedness with other conspiracy theories; The overlap of themes between anti-vaccine and other conspiracy theories creates an interdependent disinformation network, where different narratives mutually reinforce each other.
{"title":"Antivax and off-label medication communities on brazilian Telegram: between esotericism as a gateway and the monetization of false miraculous cures","authors":"Ergon Cugler de Moraes Silva","doi":"arxiv-2408.15308","DOIUrl":"https://doi.org/arxiv-2408.15308","url":null,"abstract":"Conspiracy theories, particularly those focused on anti-vaccine narratives\u0000and the promotion of off-label medications such as MMS and CDS, have\u0000proliferated on Telegram, including in Brazil, finding fertile ground among\u0000communities that share esoteric beliefs and distrust towards scientific\u0000institutions. In this context, this study seeks to answer how Brazilian\u0000conspiracy theory communities on Telegram are characterized and articulated\u0000concerning anti-vaccine themes and off-label medications? It is important to\u0000highlight that this study is part of a series of seven studies aimed at\u0000understanding and characterizing Brazilian conspiracy theory communities on\u0000Telegram. This series of seven studies is openly and originally available on\u0000the arXiv of Cornell University, applying a mirrored method across all studies,\u0000changing only the thematic object of analysis and providing replicable\u0000research, including proprietary and original codes developed, contributing to\u0000the culture of free and open-source software. Regarding the main findings of\u0000this study, it was observed: Themes such as the New World Order and Apocalypse\u0000and Survivalism act as significant gateways to anti-vaccine narratives,\u0000connecting them to theories of global control; Globalism and New World Order\u0000stand out as the main communities receiving invitations from anti-vaccine\u0000communities; Occultism and Esotericism emerge as the largest sources of\u0000invitations to off-label medication communities, creating a strong connection\u0000between esoteric beliefs and the promotion of non-scientific treatments;\u0000Anti-vaccine narratives experienced a 290% increase during the COVID-19\u0000pandemic, evidencing a growing interconnectedness with other conspiracy\u0000theories; The overlap of themes between anti-vaccine and other conspiracy\u0000theories creates an interdependent disinformation network, where different\u0000narratives mutually reinforce each other.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214924","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Conspiracy theories related to climate change denial and anti-science have found fertile ground on Telegram, particularly among Brazilian communities that distrust scientific institutions and oppose global environmental policies. This study seeks to answer the research question: how are Brazilian conspiracy theory communities on climate change and anti-science themes characterized and articulated on Telegram? It is worth noting that this study is part of a series of seven studies aimed at understanding and characterizing Brazilian conspiracy theory communities on Telegram. This series of studies is openly and originally available on arXiv from Cornell University, applying a mirrored method across all seven studies, changing only the thematic focus of analysis, and providing replicable investigation methods, including custom-developed and proprietary codes, contributing to the culture of open-source software. Regarding the main findings of this study, the following observations were made: Climate change denial and anti-science communities interact synergistically, creating a complex network that mutually reinforces disinformation narratives; Apocalyptic themes, such as Apocalypse and Survivalism, act as gateways to climate denial, with 5,057 links directed to these communities; Anti-science communities function as gatekeepers, distributing links evenly to theories such as the New World Order and Globalism, among others; During the COVID-19 pandemic, anti-science discussions experienced a significant peak, driven by vaccine disinformation; The intersection between anti-science narratives and esoteric beliefs reinforces the idea of a supposed alternative truth that challenges science; Since 2022, discussions on climate change have evolved to align with global domination theories; Additionally, the UN's 2030 Agenda is portrayed as part of a global conspiracy.
{"title":"Climate change denial and anti-science communities on brazilian Telegram: climate disinformation as a gateway to broader conspiracy networks","authors":"Ergon Cugler de Moraes Silva","doi":"arxiv-2408.15311","DOIUrl":"https://doi.org/arxiv-2408.15311","url":null,"abstract":"Conspiracy theories related to climate change denial and anti-science have\u0000found fertile ground on Telegram, particularly among Brazilian communities that\u0000distrust scientific institutions and oppose global environmental policies. This\u0000study seeks to answer the research question: how are Brazilian conspiracy\u0000theory communities on climate change and anti-science themes characterized and\u0000articulated on Telegram? It is worth noting that this study is part of a series\u0000of seven studies aimed at understanding and characterizing Brazilian conspiracy\u0000theory communities on Telegram. This series of studies is openly and originally\u0000available on arXiv from Cornell University, applying a mirrored method across\u0000all seven studies, changing only the thematic focus of analysis, and providing\u0000replicable investigation methods, including custom-developed and proprietary\u0000codes, contributing to the culture of open-source software. Regarding the main\u0000findings of this study, the following observations were made: Climate change\u0000denial and anti-science communities interact synergistically, creating a\u0000complex network that mutually reinforces disinformation narratives; Apocalyptic\u0000themes, such as Apocalypse and Survivalism, act as gateways to climate denial,\u0000with 5,057 links directed to these communities; Anti-science communities\u0000function as gatekeepers, distributing links evenly to theories such as the New\u0000World Order and Globalism, among others; During the COVID-19 pandemic,\u0000anti-science discussions experienced a significant peak, driven by vaccine\u0000disinformation; The intersection between anti-science narratives and esoteric\u0000beliefs reinforces the idea of a supposed alternative truth that challenges\u0000science; Since 2022, discussions on climate change have evolved to align with\u0000global domination theories; Additionally, the UN's 2030 Agenda is portrayed as\u0000part of a global conspiracy.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142214925","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：