Information Diffusion Patterns Research Articles

CasCIFF: A cross-domain information fusion framework tailored for cascade prediction in social networks

The explosive growth of the Internet has elevated social networks to a pivotal role in information propagation, reshaping conventional paradigms of information distribution. Utilizing the vast amount of data available online to model and predict this diffusion is of great importance in various fields. Existing approaches for information cascade prediction fall into three main categories: feature-driven methods, point process-based methods, and deep learning-based methods. Among them, deep learning-based methods, characterized by their superior learning and representation capabilities, mitigate the shortcomings inherent in the other methods.However, current deep learning methods still face several persistent challenges. In particular, accurate representation of user attributes remains problematic due to factors such as fake followers and complex network configurations. Previous algorithms that focused on the sequential order of user activations often neglected the rich insights offered by activation timing. Furthermore, these techniques often fail to holistically integrate temporal and structural aspects, thus missing the nuanced propagation trends inherent in information cascades.To address these issues, we propose the Cross-Domain Information Fusion Framework (CasCIFF), which is tailored for information cascade prediction. This framework exploits multi-hop neighborhood information to make user embeddings robust. When embedding cascades, the framework intentionally incorporates timestamps, endowing it with the ability to capture evolving patterns of information diffusion. In particular, CasCIFF seamlessly integrates the tasks of user classification and cascade prediction into a consolidated framework, thereby allowing the extraction of common features that prove useful for all tasks, a strategy anchored in the principles of multi-task learning. After extensive experiments conducted on publicly available datasets, the results demonstrate CasCIFF's superiority over established baseline methods in terms of prediction accuracy.

Differentiating broadcast from viral: a causal inference approach for information diffusion analysis

Classifying information diffusion patterns is critical to many information analysis areas, e.g., misleading information detection. However, diffusion pattern classification remains challenging when multiple users are involved. To address this challenge, this study aims to classify how information diffuses, distinguishing between broadcast and viral spreading, solely through the analysis of observational data from retweet networks on X (formerly known as Twitter). In broadcasting, most users directly receive information. However, viral spreading allows users the opportunity to receive information from a variety of sources. Therefore, viral spreading increases the likelihood of identifying misleading information. Existing methods classify diffusion types mainly through structural virality, which relies on the average distance between the users. However, when dealing with diffusion networks involving two or more information sources, these approaches can potentially lead to confusion regarding causality. To tackle this problem, we develop a deterministic causal inference method for categorizing information diffusion types. To the best of our knowledge, this is the first study investigating information diffusion types based on causality. This approach can be used to assess source credibility and assist in detecting misleading information. It can also be extended to other social media platforms.Graphical

Account credibility inference based on news-sharing networks

The spread of misinformation poses a threat to the social media ecosystem. Effective countermeasures to mitigate this threat require that social media platforms be able to accurately detect low-credibility accounts even before the content they share can be classified as misinformation. Here we present methods to infer account credibility from information diffusion patterns, in particular leveraging two networks: the reshare network, capturing an account’s trust in other accounts, and the bipartite account-source network, capturing an account’s trust in media sources. We extend network centrality measures and graph embedding techniques, systematically comparing these algorithms on data from diverse contexts and social media platforms. We demonstrate that both kinds of trust networks provide useful signals for estimating account credibility. Some of the proposed methods yield high accuracy, providing promising solutions to promote the dissemination of reliable information in online communities. Two kinds of homophily emerge from our results: accounts tend to have similar credibility if they reshare each other’s content or share content from similar sources. Our methodology invites further investigation into the relationship between accounts and news sources to better characterize misinformation spreaders.

Lessons Learned From Interdisciplinary Efforts to Combat COVID-19 Misinformation: Development of Agile Integrative Methods From Behavioral Science, Data Science, and Implementation Science.

Despite increasing awareness about and advances in addressing social media misinformation, the free flow of false COVID-19 information has continued, affecting individuals' preventive behaviors, including masking, testing, and vaccine uptake. In this paper, we describe our multidisciplinary efforts with a specific focus on methods to (1) gather community needs, (2) develop interventions, and (3) conduct large-scale agile and rapid community assessments to examine and combat COVID-19 misinformation. We used the Intervention Mapping framework to perform community needs assessment and develop theory-informed interventions. To supplement these rapid and responsive efforts through large-scale online social listening, we developed a novel methodological framework, comprising qualitative inquiry, computational methods, and quantitative network models to analyze publicly available social media data sets to model content-specific misinformation dynamics and guide content tailoring efforts. As part of community needs assessment, we conducted 11 semistructured interviews, 4 listening sessions, and 3 focus groups with community scientists. Further, we used our data repository with 416,927 COVID-19 social media posts to gather information diffusion patterns through digital channels. Our results from community needs assessment revealed the complex intertwining of personal, cultural, and social influences of misinformation on individual behaviors and engagement. Our social media interventions resulted in limited community engagement and indicated the need for consumer advocacy and influencer recruitment. The linking of theoretical constructs underlying health behaviors to COVID-19-related social media interactions through semantic and syntactic features using our computational models has revealed frequent interaction typologies in factual and misleading COVID-19 posts and indicated significant differences in network metrics such as degree. The performance of our deep learning classifiers was reasonable, with an F-measure of 0.80 for speech acts and 0.81 for behavior constructs. Our study highlights the strengths of community-based field studies and emphasizes the utility of large-scale social media data sets in enabling rapid intervention tailoring to adapt grassroots community interventions to thwart misinformation seeding and spread among minority communities. Implications for consumer advocacy, data governance, and industry incentives are discussed for the sustainable role of social media solutions in public health.

Exploring Spillover Effects for COVID-19 Cascade Prediction.

An information outbreak occurs on social media along with the COVID-19 pandemic and leads to an infodemic. Predicting the popularity of online content, known as cascade prediction, allows for not only catching in advance information that deserves attention, but also identifying false information that will widely spread and require quick response to mitigate its negative impact. Among the various information diffusion patterns leveraged in previous works, the spillover effect of the information exposed to users on their decisions to participate in diffusing certain information has not been studied. In this paper, we focus on the diffusion of information related to COVID-19 preventive measures due to its special role in consolidating public efforts to slow down the spread of the virus. Through our collected Twitter dataset, we validate the existence of the spillover effects. Building on this finding, we propose extensions to three cascade prediction methods based on Graph Neural Networks (GNNs). Experiments conducted on our dataset demonstrated that the use of the identified spillover effects significantly improves the state-of-the-art GNN methods in predicting the popularity of not only preventive measure messages, but also other COVID-19 messages.

Uncertainty matters in US financial information spillovers: Evidence from a directed acyclic graph approach

We investigate the role of uncertainty in US financial information spillovers by introducing uncertainty measures into financial market networks from January 1986 to April 2020. A directed acyclic graph (DAG) approach is utilized to determine the network structure and reveal the contemporaneous causalities among financial markets and uncertainty. We empirically show that uncertainty matters in US financial information spillovers, and it behaves differently between return and volatility spillovers. News-based implied volatility (NVIX), capturing investors’ perception of future uncertainty, is an important linkage of financial returns, acting as an exogenous shock to financial volatility spillovers. A subsample analysis indicates that financial uncertainty (FU) matters before the financial crisis, and its role is replaced by NVIX after the crisis. We provide a new perspective to understand the information diffusion patterns of US financial markets, which has meaningful implications for investors who adjust their investment allocations according to spillovers and for policymakers who could rely on NVIX as an important indicator of financial contagions.

Public Engagement in Climate Communication on China’s Weibo: Network Structure and Information Flows

This article provides an empirical study of public engagement with climate change discourse in China by analysing how Chinese publics participate in the public discussion around two Intergovernmental Panel on Climate Change reports and how individual users interact with state and elite actors on the pre-eminent Chinese microblogging platform Weibo. Using social network analysis methods and a temporal comparison, we examine the structure of climate communication networks, the direction of information flows among multiple types of Weibo users, and the changes in information diffusion patterns between the pre- and post-Paris periods. Our results show there is an increasing yet constrained form of public engagement in climate communication on Weibo alongside China’s pro-environmental transition in recent years. We find an expansion of public engagement as shown by individual users’ increasing influence in communication networks and the diversification of frames associated with climate change discourse. However, we also find three restrictive interaction tendencies that limit Weibo’s potential to facilitate multi-directional communication and open public deliberation of climate change, including the decline of mutually balanced dialogic interactions, the lack of bottom-up information flows, and the reinforcement of homophily tendencies amongst eco-insiders and governmental users. These findings highlight the coexistence of both opportunities and constraints of Weibo being a venue for public engagement with climate communication and as a forum for a new climate politics and citizen participation in China.

Uncovering information diffusion patterns in different networks using the L-metric

ABSTRACT Information diffusion is an important branch of online social network analysis. In this paper, we construct a new metric, the proportion of leaf nodes in a diffusion tree (L-metric), to quantify information diffusion patterns, and we study the impact of the network category and information content on these patterns. Simulation-based experimental studies of real-world social networks show that information diffusion exhibits different patterns in different networks, and niche information does not typically propagate easily in any type of network. These conclusions provide a new perspective for further research on management decisions with regard to online social networks.

Network Representation Learning: A Survey

With the widespread use of information technologies, information networks are becoming increasingly popular to capture complex relationships across various disciplines, such as social networks, citation networks, telecommunication networks, and biological networks. Analyzing these networks sheds light on different aspects of social life such as the structure of societies, information diffusion, and communication patterns. In reality, however, the large scale of information networks often makes network analytic tasks computationally expensive or intractable. Network representation learning has been recently proposed as a new learning paradigm to embed network vertices into a low-dimensional vector space, by preserving network topology structure, vertex content, and other side information. This facilitates the original network to be easily handled in the new vector space for further analysis. In this survey, we perform a comprehensive review of the current literature on network representation learning in the data mining and machine learning field. We propose new taxonomies to categorize and summarize the state-of-the-art network representation learning techniques according to the underlying learning mechanisms, the network information intended to preserve, as well as the algorithmic designs and methodologies. We summarize evaluation protocols used for validating network representation learning including published benchmark datasets, evaluation methods, and open source algorithms. We also perform empirical studies to compare the performance of representative algorithms on common datasets, and analyze their computational complexity. Finally, we suggest promising research directions to facilitate future study.

Crowd or Hubs: information diffusion patterns in online social networks in disasters

The objective of this paper is to investigate the role of different types of users in the diffusion of situational information through online social networks in disasters. In particular, this paper investigates the influence of two types of users: crowd (regular users) and hubs (users with a large number of followers) on the speed and magnitude of information propagation. Effective and efficient disaster response requires rapid dissemination of situational information to improve situation awareness, save lives, and quickly repair damages. The use of social media, such as Twitter, has gained popularity for spreading the situational information in disasters. Little is known, however, regarding the underlying diffusion mechanisms influencing the speed and magnitude of spreading disaster-related information on social media. To address this gap, using tweets related to Hurricane Harvey, this study examined the role of hubs and crowd and the influence of different features such as theme, hashtag, media, users’ location, and the intervention timing of online users on the speed and magnitude of information spread. The results compare the differences in the speed and magnitude of information spread between two diffusion patterns: crowd diffusion (less than 2% retweets from hubs) and mixed diffusion (more than 2% retweets from hubs). The findings suggest that both diffusion patterns can achieve high speed and high magnitude in terms of information diffusion for trending tweets with different features. For mixed diffusion, the speed and magnitude of the tweets are governed by the activities of hubs. The results also show that early intervention of hubs increases the speed of information propagation. Also, in the crowd diffusion, information spread is influenced by both crowd and hubs whose retweets cause tipping points in the information diffusion process at different points of time. The findings imply intervention strategies to better disseminate situational information in disasters.

Uncertainty-based False Information Propagation in Social Networks

Many network scientists have investigated the problem of mitigating or removing false information propagated in social networks. False information falls into two broad categories: disinformation and misinformation. Disinformation represents false information that is knowingly shared and distributed with malicious intent. Misinformation in contrast is false information shared unwittingly, without any malicious intent. Many existing methods to mitigate or remove false information in networks concentrate on methods to find a set of seeding nodes (or agents) based on their network characteristics (e.g., centrality features) to treat. The aim of these methods is to disseminate correct information in the most efficient way. However, little work has focused on the role of uncertainty as a factor in the formulation of agents’ opinions. Uncertainty-aware agents can form different opinions and eventual beliefs about true or false information resulting in different patterns of information diffusion in networks. In this work, we leverage an opinion model, called Subjective Logic (SL), which explicitly deals with a level of uncertainty in an opinion where the opinion is defined as a combination of belief, disbelief, and uncertainty, and the level of uncertainty is easily interpreted as a person’s confidence in the given belief or disbelief. However, SL considers the dimension of uncertainty only derived from a lack of information (i.e., ignorance), not from other causes, such as conflicting evidence. In the era of Big Data, where we are flooded with information, conflicting information can increase uncertainty (or ambiguity) and have a greater effect on opinions than a lack of information (or ignorance). To enhance the capability of SL to deal with ambiguity as well as ignorance, we propose an SL-based opinion model that includes a level of uncertainty derived from both causes. By developing a variant of the Susceptible-Infected-Recovered epidemic model that can change an agent’s status based on the state of their opinions, we capture the evolution of agents’ opinions over time. We present an analysis and discussion of critical changes in network outcomes under varying values of key design parameters, including the frequency ratio of true or false information propagation, centrality metrics used for selecting seeding false informers and true informers, an opinion decay factor, the degree of agents’ prior belief, and the percentage of true informers. We validated our proposed opinion model using both the synthetic network environments and realistic network environments considering a real network topology, user behaviors, and the quality of news articles. The proposed agent’s opinion model and corresponding strategies to deal with false information can be applicable to combat the spread of fake news in various social media platforms (e.g., Facebook).

Space, time, and disease on social media: a case study of dengue fever in China

It is possible to generate real-time and location-by-location data of many types of human dynamic events based on social media information for the awareness of events in public health. Analyzing these events is useful in understanding spatiotemporal trends and patterns of how diseases spread and also provides indications for users’ sentiment about the concerned disease. This article examines the spatial and temporal patterns of social media posts based on the content, attributes, and follower activities of posts on social media. We describe the spatial features of the topic discussed in the posts and the spatial relationship among comments on the posts. We present models for describing the diffusion process of these posts and for exploring their spatiotemporal patterns. Our results suggest that (1) the long-term trends of the topics in users’ views seem to be stable, (2) results from analyzing follower activities of posts are critical in describing the spatial patterns of the posts, and (3) the diffusion process of an event in social media is still similar to that of a traditional information diffusion model. Our findings are useful for understanding social media and social events. The processes we describe in this article suggest a standard form of analysis that can be adopted for extracting spatiotemporal patterns of information diffusion and for data mining in social media posts.

Partial differential equation modeling with Dirichlet boundary conditions on social networks

The being a wide range of applications of the Internet, social networks have become an effective and convenient platform for information communication, propagation and diffusion. Most of information exchange and spreading exist in social networks. The issue of information diffusion in social networks is getting more and more attention by government and individuals. The researchers investigated either empirical studies or focused on ordinary differential equation (ODE) models with only consideration of temporal dimension in most prior work. As is well known, partial differential equations (PDEs) can describe temporal and spatial patterns of information diffusion over online social networks; however, until now, results for understanding information propagation of social networks over both temporal and spatial dimensions are few. This paper is devoted to investigating a non-autonomous diffusive logistic model with Dirichlet boundary conditions to describe the process of information propagation in social networks. By constructing upper and lower solutions we obtain the dynamic behavior of the solution to the non-autonomous diffusive logistic model. Our results show that information diffusion is greatly affected by the diffusion coefficient d(t) and the intrinsic growth rate r(t).

Influential Users in Social Network Services

Social network services (SNSs) are fundamentally transforming the way businesses communicate with customers. Key issues in understanding social media and such services include how to generate more information traffic and how to increase information exposure. Unlike previous studies in the research stream studying patterns of information diffusion, we investigate the role of influential Twitter users, focusing on their positions in social networks, in redistributing contents generated by other users. To do this, we employ sociological theories in determining which characteristics of connecting users provide more influence. Specifically, a connecting user's position in networks observable to audience affects the user's influence in a way that connecting users with relatively high status and those embedded in a closely connected community exert the greatest impact on the network. Moreover, the value of a connecting user's network position can be contingent on the source user's status and network position: a high-status connecting user contributes more to information diffusion when tweets are generated by a low-status source user, and a connecting user who is embedded in a well-connected network contributes more to information diffusion when tweets are generated by a source user in a brokerage position. After exploring data on tweets about Apple products and services, we find evidence that support this set of theoretical claims. Such findings would also provide a useful guidance for practitioners to enhance information diffusion by effectively designing the sequence of content recipients based on their network positions.

Modeling the Impact of Innovation Diffusion on Solar PV Adoption in City Neighborhoods

We present an agent-based model for Renewable Energy Technologies (RET) in Qatar city neighborhoods based on the spread of information in social networks within city-neighborhoods. Information diffusion patterns in household and Twitter networks are combined to model the rate of diffusion of RET innovation. The resulting approach provides a methodology for capturing how RET innovation diffusion in online social networks and city neighborhood networks may jointly influence the residential adoption of renewable energy technologies. We demonstrate an application of this approach to an agent-based model of solar PV adoption in Qatar.

Nonlinear Dynamics of Information Diffusion in Social Networks

The recent explosion in the adoption of search engines and new media such as blogs and Twitter have facilitated the faster propagation of news and rumors. How quickly does a piece of news spread over these media? How does its popularity diminish over time? Does the rising and falling pattern follow a simple universal law? In this article, we propose SpikeM, a concise yet flexible analytical model of the rise and fall patterns of information diffusion. Our model has the following advantages. First, unification power: it explains earlier empirical observations and generalizes theoretical models including the SI and SIR models. We provide the threshold of the take-off versus die-out conditions for SpikeM and discuss the generality of our model by applying it to an arbitrary graph topology. Second, practicality: it matches the observed behavior of diverse sets of real data. Third, parsimony: it requires only a handful of parameters. Fourth, usefulness: it makes it possible to perform analytic tasks such as forecasting, spotting anomalies, and interpretation by reverse engineering the system parameters of interest (quality of news, number of interested bloggers, etc.). We also introduce an efficient and effective algorithm for the real-time monitoring of information diffusion, namely SpikeStream, which identifies multiple diffusion patterns in a large collection of online event streams. Extensive experiments on real datasets demonstrate that SpikeM accurately and succinctly describes all patterns of the rise and fall spikes in social networks.

Construct Demo Input Deck

Abstract : This technical report describes a demo input deck for Construct, the CASOS dynamic network simulation tool. It describes an example problem - understanding the difference in information diffusion patterns when bridging agents span two otherwise separate groups - and then how such a simulation would be simulated using Construct. In so doing, it describes the way in which data is supplied to Construct. Additionally, this report also describes how to run the simulation using the command line, and ways in which the simulation could be extended to analyze other problems.

A Study on the Activities of Touch Point and Consumer Information Search in the Multimedia Environment

The purpose of this study is, in the situation that consumers use not only online news but social media as information delivery media, based on web data, to examine the effect of touch point activities in two media different according to product types on consumer search activities. To do this, we traced the patterns of information diffusion according to time, by extracting the number of the web postings of the information about the products from on-line news and blogs on a weekly or monthly basis, and identified consumer search activities about the goods by using their search traffics provided on a weekly basis by portal sites As a result of the study, it turned out that touch point activities and consumer search activities are different from each other in two media different according to product types. Specifically speaking, as for utilitarian goods, it appeared that there were correlations between the touch points and consumer search activities of the two different information media. In particular, more similarity in the patterns of information diffusion was found in the news sites than in the blogs. In the case of hedonic goods, the correlations between touch points and consumer search activities showed different results according to the time intervals. A monthly analysis showed that the consumer search activities had correlations not with on-line news, but with blogs. On the contrary, a weekly analysis showed that they had higher correlations with blogs than with on-line news sites though having correlations with both the two.

Analysis of micro-blog diffusion using a dynamic fluid model

Various methods on the display of dynamic information diffusion for social media have been proposed. Most of them use data mining approaches to explore the behaviors and interactions between users. Such approaches are unable to reveal the complex mechanism and the process of information diffusion. Lattice Boltzmann Method (LBM) models fluid behaviors at the microscopic scale, similar to the information diffusion in social media that is determined by the collective behavior of many personal retweeting of topics. We propose an information diffusion model inspired by the fundamental idea of LBM to analyze and simulate users’ communicating behaviors and processes in Micro-blogging. The micro-blog space is regarded as an artificial physical system with social phenomena such as micro-blog bursting. The macroscopic properties of the information diffusion model are explored to simulate and predict the trend of information diffusion for any specific topic. A novel visualization style mimicking fluid dynamics is proposed to help understand the scale of information diffusion and the popularity of a topic. The flow visualization based on the speed of information diffusion is useful in discovering typical information diffusion patterns for different types of topics in social networks. Comparing with other approaches, our approach provides more effective yet intuitive simulation.

Partial differential equations with Robin boundary condition in online social networks

In recent years, online social networks such as Twitter, have become a major source of information exchange and research on information diffusion in social networks has been accelerated. Partial differential equations are proposed to characterize temporal and spatial patterns of information diffusion over online social networks. The new modeling approach presents a new analytic framework towards quantifying information diffusion through the interplay of structural and topical influences. In this paper we develop a non-autonomous diffusive logistic model with indefinite weight and the Robin boundary condition to describe information diffusion in online social networks. It is validated with a real dataset from an online social network, Digg.com. The simulation shows that the logistic model with the Robin boundary condition is able to more accurately predict the density of influenced users. We study the bifurcation, stability of the diffusive logistic model with heterogeneity in distance. The bifurcation and stability results of the model information describe either information spreading or vanishing in online social networks.