Rumor Verification Research Articles

Rumor gatekeepers: Unsupervised ranking of Arabic twitter authorities for information verification

The advent of online social networks (OSNs) has catalyzed the formation of novel learning communities. Identifying experts within OSNs has become a critical component for facilitating knowledge exchange and enhancing self-awareness, particularly in contexts such as rumor verification processes. Research efforts aimed at locating authorities in OSNs are scant, largely due to the scarcity of annotated datasets. This work represents a contribution to the domain of unsupervised learning to address the challenge of authorities’ identification in Twitter. We have employed advanced natural language processing technique to transfer knowledge concerning topics in the Arabic language and to discern the semantic connections among candidates within Twitter in zero-shot learning. We take advantage of the Single-labeled Arabic News Articles Dataset (SANAD) to perform the process of extracting domain features and applying these features in finding authorities using the Authority Finding in Arabic Twitter (AuFIN) dataset. Our evaluation assessed the extent of extracted topical features transferred and the efficacy of authorities’ retrieval in comparison to the latest unsupervised models in this domain. Our approach successfully extracted and integrated the limited available topical semantic features of the language into the representation of candidates. The findings indicate that our hybrid model surpasses those that rely solely on lexical features of language and network topology, as well as other contemporary approaches to topic-specific expert finding.

Filter-based Stance Network for Rumor Verification

Rumor verification on social media aims to identify the truth value of a rumor, which is important to decrease the detrimental public effects. A rumor might arouse heated discussions and replies, conveying different stances of users that could be helpful in identifying the rumor. Thus, several works have been proposed to verify a rumor by modelling its entire stance sequence in the time domain. However, these works ignore that such a stance sequence could be decomposed into controversies with different intensities, which could be used to cluster the stance sequences with the same consensus. In addition, the existing stance extractors fail to consider both the impact of all previously posted tweets and the reply chain on obtaining the stance of a new reply. To address the above problems, in this article, we propose a novel stance-based network to aggregate the controversies of the stance sequence for rumor verification, termed Filter-based Stance Network (FSNet). As controversies with different intensities are reflected as the different changes of stances, it is convenient to represent different controversies in the frequency domain, but it is hard in the time domain. Our proposed FSNet decomposes the stance sequence into multiple controversies in the frequency domain and obtains the weighted aggregation of them. Specifically, FSNet consists of two modules: the stance extractor and the filter block. To obtain better stance features toward the source, the stance extractor contains two stages. In the first stage, the tweet representation of each reply is obtained by aggregating information from all previously posted tweets in a conversation. Then, the features of stance toward the source, i.e., rumor-aware stance, are extracted with the reply chains in the second stage. In the filter block module, a rumor-aware stance sequence is constructed by sorting all the tweets of a conversation in chronological order. Fourier Transform thereafter is employed to convert the stance sequence into the frequency domain, where different frequency components reflect controversies of different intensities. Finally, a frequency filter is applied to explore the different contributions of controversies. We supervise our FSNet with both stance labels and rumor labels to strengthen the relations between rumor veracity and crowd stances. Extensive experiments on two benchmark datasets demonstrate that our model substantially outperforms all the baselines.

Synews: a synergy-based rumor verification system

Synews: a synergy-based rumor verification system

SSRI-Net: Subthreads Stance–Rumor Interaction Network for rumor verification

As online rumors have the potential to greatly affect areas such as social order, stock prices, and presidential elections, there is an emerging necessity for the automation of rumor verification. Although the current methods have achieved satisfactory performance, they still suffer from the following problems. First, the current methods simply concatenate the representations of different subthreads in their models, which may result in omitting some important information. Second, although stance information has been considered for the rumor verification task, it has not been fully utilized. To solve the problems, we propose the Subthreads Stance–Rumor Interaction Network (SSRI-Net) model for rumor verification. The proposed SSRI-Net model first introduces the Subthreads Interaction Attention mechanism between different subthreads to capture the interaction information between subthreads for a better understanding on user posts. Moreover, we also design the Stance–Rumor Interaction Network to fully integrate users’ stance information with rumor verification. We have conducted experiments on two public datasets, namely SemEval-2017 and PHEME datasets, for performance evaluation. Our SSRI-Net model outperforms the previous best models by 5.8% and 7.1% in Macro-F1 and Accuracy respectively on the SemEval-2017 dataset. In addition, our SSRI-Net model also outperforms the previous best models by 4.7% and 5.4% in Macro-F1 and Accuracy respectively on the PHEME dataset. The experimental results have shown that our proposed SSRI-Net model has outperformed the baseline models and achieved the state-of-the-art performance for rumor verification.

Are authorities denying or supporting? Detecting stance of authorities towards rumors in Twitter

Several studies examined the leverage of the stance in conversational threads or news articles as a signal for rumor verification. However, none of these studies leveraged the stance of trusted authorities. In this work, we define the task of detecting the stance of authorities towards rumors in Twitter, i.e., whether a tweet from an authority supports the rumor, denies it, or neither. We believe the task is useful to augment the sources of evidence exploited by existing rumor verification models. We construct and release the first Authority STance towards Rumors (AuSTR) dataset, where evidence is retrieved from authority timelines in Arabic Twitter. The collection comprises 811 (rumor tweet, authority tweet) pairs relevant to 292 unique rumors. Due to the relatively limited size of our dataset, we explore the adequacy of existing Arabic datasets of stance towards claims in training BERT-based models for our task, and the effect of augmenting AuSTR with those datasets. Our experiments show that, despite its limited size, a model trained solely on AuSTR with a class-balanced focus loss exhibits a comparable performance to the best studied combination of existing datasets augmented with AuSTR, achieving a performance of 0.84 macro-F1 and 0.78 F1 on debunking tweets. The results indicate that AuSTR can be sufficient for our task without the need for augmenting it with existing stance datasets. Finally, we conduct a thorough failure analysis to gain insights for the future directions on the task.

Causal diffused graph-transformer network with stacked early classification loss for efficient stream classification of rumours

The growth in social media has led to the increasing spread of unverified or false information. Automatically detecting rumours and accessing their veracity, i.e., false rumours, true rumours, or unverified rumours, is an important and challenging task in social media analytics. This paper aims to build an effective and scalable stream classification framework for early fine-grained rumour classification based on community response. We propose a Causal Diffused Graph-Transformer Network (CDGTN) to extract features from the source-reply graph in a social media conversation. Then, we propose Source-Guided Incremental Attention Pooling (SGIAP) to aggregate the encoded features with discrete timestamps. To improve the performance of early classification, we propose a Stacked Early Classification Loss (SecLoss), which aims to minimize the classification loss over the time instances. This can greatly improve the effectiveness of early classification of rumours. To improve the efficiency of streaming rumour verification, we propose a continued inference algorithm based on prefix-sum, which can greatly reduce the computational complexity of stream classification of rumours. Furthermore, we annotated the first Chinese rumour verification dataset, by extending the existing Chinese-Twitter dataset, namely CR-Twitter, originally for rumour detection. We conducted experiments on the Twitter15, Twitter16, PHEME, Weibo, and the extended CR-Twitter datasets for rumour classification, to verify our proposed stream classification framework. The experimental results show that our proposed framework can significantly boost the effectiveness and efficiency of early stream classification of rumours. Models and datasets are released at: https://thcheung.github.io/cdgtn/.

Who can verify this? Finding authorities for rumor verification in Twitter

A large body of research work has proposed verification techniques for rumors spreading in social media that mainly relied on subjective evidence, e.g., propagation networks or user interactions. Alternatively, in this work, we introduce the task of authority finding in social media, in which we aim to find authorities, for given rumors spreading specifically in Twitter, who can help verify them by providing exclusive/convincing evidence that supports or denies those rumors. We release the first test collection for Authority FINding in Arabic Twitter (AuFIN). The collection comprises 150 rumors (expressed in tweets) associated with a total of 1,044 authority accounts and a user collection of 395,231 Twitter accounts (members of 1,192,284 unique Twitter lists). Moreover, we propose a hybrid model that employs pre-trained language models and combines lexical, semantic, and network signals to find authorities. Our experiments show that the textual representation of users is insufficient, and incorporating the Twitter network features improved the recall of authorities by 34%. Moreover, semantic ranking is inferior to the lexical and network-based ranking in terms of precision, but superior in terms of recall. Therefore, combining both the semantic and network-based ranking achieved the best overall performance achieving a precision of 0.413 and 0.213 at depth 1 and 5 respectively. We show that rumor expansion by exploiting Knowledge Bases improves the recall of authorities by up to 15%. Furthermore, we find that SOTA models for topic expert finding perform poorly on finding authorities. Finally, drawing upon our experiments, we discuss failure factors and make recommendations for future research directions in addressing this task.

Review of stance detection for rumor verification in social media

Social media is a perfect breeding ground for false rumors due to the simplicity of sharing information, which may have negative implications in a variety of domains, including economics, healthcare, and politics. Previous research indicates that public reactions to false rumors are a critical indicator for determining the truthfulness of news. In social media, substantial effort has been invested in detecting and debunking rumors based on crowd stance, given that stance is a vital part of automatic news verification. This paper presents a review of recent approaches in the area of rumor verification using stance detection, which attempts to determine a given document’s stance with respect to a given piece of news. The review offers a detailed list of datasets, as well as a summary of relevant experiments and methods employed, as well as analysis of helpful features for addressing this issue. Finally, we highlight the main challenges and future directions in this field by utilizing stance detection.

Evaluating the generalisability of neural rumour verification models

Research on automated social media rumour verification, the task of identifying the veracity of questionable information circulating on social media, has yielded neural models achieving high performance, with accuracy scores that often exceed 90%. However, none of these studies focus on the real-world generalisability of the proposed approaches, that is whether the models perform well on datasets other than those on which they were initially trained and tested. In this work we aim to fill this gap by assessing the generalisability of top performing neural rumour verification models covering a range of different architectures from the perspectives of both topic and temporal robustness. For a more complete evaluation of generalisability, we collect and release COVID-RV, a novel dataset of Twitter conversations revolving around COVID-19 rumours. Unlike other existing COVID-19 datasets, our COVID-RV contains conversations around rumours that follow the format of prominent rumour verification benchmarks, while being different from them in terms of topic and time scale, thus allowing better assessment of the temporal robustness of the models. We evaluate model performance on COVID-RV and three popular rumour verification datasets to understand limitations and advantages of different model architectures, training datasets and evaluation scenarios. We find a dramatic drop in performance when testing models on a different dataset from that used for training. Further, we evaluate the ability of models to generalise in a few-shot learning setup, as well as when word embeddings are updated with the vocabulary of a new, unseen rumour. Drawing upon our experiments we discuss challenges and make recommendations for future research directions in addressing this important problem.

Interaction between official institutions and influential users of rumor control in online social networks.

Online interactions have become major channels for people to obtain and disseminate information during the new normal of COVID-19, which can also be a primary platform for rumor propagation. There are many complex psychological reasons for spreading rumors, but previous studies have not fully analyzed this problem from the perspective of the interaction between official institutions and influential users. The purpose of this study is to determine optimal strategies for official institutions considering the impact of two different influential user types (trolls and reputed personalities) by designing two game-theoretic models, namely “Rumor Clarification and Interaction Model” and “Rumor Verification and Interaction Model,” which can, respectively decide whether to clarify and when to clarify. The results of this article show that clarification strategies can be decided according to the characteristics of rumors and the influential user’s reactions. Meanwhile, publishing verified information prevents trolls’ “loophole advantages” and prevents reputed personalities from spreading false information due to the vague authenticity of rumors. Results also show that the verification strategy is limited by cost, period, and verification index.

A model to measure the spread power of rumors.

With technologies that have democratized the production and reproduction of information, a significant portion of daily interacted posts in social media has been infected by rumors. Despite the extensive research on rumor detection and verification, so far, the problem of calculating the spread power of rumors has not been considered. To address this research gap, the present study seeks a model to calculate the Spread Power of Rumor (SPR) as the function of content-based features in two categories: False Rumor (FR) and True Rumor (TR). For this purpose, the theory of Allport and Postman will be adopted, which it claims that importance and ambiguity are the key variables in rumor-mongering and the power of rumor. Totally 42 content features in two categories “importance” (28 features) and “ambiguity” (14 features) are introduced to compute SPR. The proposed model is evaluated on two datasets, Twitter and Telegram. The results showed that (i) the spread power of False Rumor documents is rarely more than True Rumors. (ii) there is a significant difference between the SPR means of two groups False Rumor and True Rumor. (iii) SPR as a criterion can have a positive impact on distinguishing False Rumors and True Rumors.

GDART: Improving rumor verification in social media with Discrete Attention Representations

Due to the harmful impact of fabricated information on social media, many rumor verification techniques have been introduced in recent years. Advanced techniques like multi-task learning (MTL), shared-private models suffer from many strategic limitations that restrict their capability of veracity identification on social media. These models are often reliant on multiple tasks for the primary targeted objective. Even the most recent deep neural network (DNN) models like VRoC, Hierarchical-PSV, StA-HiTPLAN etc. based on VAE, GCN, Transformer respectively with improved modification are able to perform good on veracity identification task but with the help of additional auxiliary information, mostly. However, their rise is still not substantial with respect to the proposed model even though the proposed model is not using any additional information. To come up with an improved DNN model architecture, we introduce globally Discrete Attention Representations from Transformers (gDART). Discrete-Attention mechanism in gDART is capable of capturing multifarious correlations veiled among the sequence of words which existing DNN models including Transformer often overlook. Our proposed framework uses a Branch-CoRR Attention Network to extract highly informative features in branches, and employs Feature Fusion Network Component to identify deep embedded features and use them to make enhanced identification of veracity of an unverified claim. Moreover, to achieve its goal, gDART is not dependent on any costly auxiliary resource but on an unsupervised learning process. Extensive experiments reveal that gDART marks a considerable performance gain in veracity identification task over state-of-the-art models on two real world rumor datasets. gDART reports a gain of 36.76%, 40.85% on standard benchmark metrics.

Rumor Stance Classification in Online Social Networks: The State-of-the-Art, Prospects, and Future Challenges

The emergence of the Internet as a ubiquitous technology has facilitated the rapid evolution of social media as the leading virtual platform for communication, content sharing, and information dissemination. In spite of revolutionizing the way news is delivered to people, this technology has also brought along with itself inevitable demerits. One such drawback is the spread of rumors expedited by social media platforms, which may provoke doubt and fear. Therefore, it is essential to debunk rumors before their widespread use. Over the years, many studies have been conducted to develop effective rumor verification systems. One aspect of such studies focuses on rumor stance classification, which involves the task of utilizing user viewpoints regarding a rumorous post to better predict the veracity of a rumor. Relying on user stances in rumor verification has gained significant importance, for it has resulted in significant improvements in the model performance. In this paper, we conduct a comprehensive literature review of rumor stance classification in complex online social networks (OSNs). In particular, we present a thorough description of these approaches and compare their performances. Moreover, we introduce multiple datasets available for this purpose and highlight their limitations. Finally, challenges and future directions are discussed to stimulate further relevant research efforts.

A Deep Content-Based Model for Persian Rumor Verification

During the development of social media, there has been a transformation in social communication. Despite their positive applications in social interactions and news spread, it also provides an ideal platform for spreading rumors. Rumors can endanger the security of society in normal or critical situations. Therefore, it is important to detect and verify the rumors in the early stage of their spreading. Many research works have focused on social attributes in the social network to solve the problem of rumor detection and verification, while less attention has been paid to content features. The social and structural features of rumors develop over time and are not available in the early stage of rumor. Therefore, this study presented a content-based model to verify the Persian rumors on Twitter and Telegram early. The proposed model demonstrates the important role of content in spreading rumors and generates a better-integrated representation for each source rumor document by fusing its semantic, pragmatic, and syntactic information. First, contextual word embeddings of the source rumor are generated by a hybrid model based on ParsBERT and parallel CapsNets. Then, pragmatic and syntactic features of the rumor are extracted and concatenated with embeddings to capture the rich information for rumor verification. Experimental results on real-world datasets demonstrated that the proposed model significantly outperforms the state-of-the-art models in the early rumor verification task. Also, it can enhance the performance of the classifier from 2% to 11% on Twitter and from 5% to 23% on Telegram. These results validate the model's effectiveness when limited content information is available.

Rumor Identification and Verification for Text in Social Media Content

Abstract Twitter led a remarkable breakthrough in information sharing on online social media. The eminent technology can propagate a piece of rumor to a large community of people in a short period. The timely detection of rumor tweets in social media curtails panic among the public during critical situations. Traditional machine learning techniques are not capable of categorizing rumor information effectively. To address this problem, the author has proposed a novel neural network approach called veracity detection neural network for identifying the rumor-related Twitter posts’ content in real-time events. This algorithm utilized the convolutional sentence encoder–bi-directional long short-term memory (CSE-BiLSTM) model with pre-trained vectorization methods such as Word2vec, fastText and universal sentence encoder (USE). The hybrid CSE-BiLSTM with USE vectorization technique yields the best results for the performance metrics of accuracy, F1-score, precision and recall. The proposed algorithm achieves 90.56%, 86.18% and 93.89% accuracy values to classify the tweet into rumor or non-rumor for the datasets such as PHEME, newly emerged rumors on Twitter and #gaja, respectively. Finally, a comparative study shows that the proposed neural network model outperformed all other existing rumor text classification systems.

A semi-supervised model for Persian rumor verification based on content information

Rumor is a collective attempt to interpret a vague but attractive situation by using the power of words. In social networks, false-rumors may have significantly different contextual characteristics from true-rumors at lexical, syntactic, semantic levels. Therefore, this study presents the BERT-SAWS semi-supervised learning model for early verification of Persian rumor by investigating content-based and context features at three views: Contextual Word Embeddings (CWE), speech act, and Writing Style (WS). This model is built by loading pre-trained Bidirectional Encoder Representations from Transformers (BERT) as an unsupervised language representation, fine-tuning it using a small Persian rumor dataset, and combining with a supervised learning model to provide an enriched text representation of the content of the rumor. This text representation enables the model to have a better comprehending of the rumor language to verify rumors better than baseline models for two reasons: (i) early rumor verification by focusing on content-based and context-based features of the source rumor. (ii) overcoming the problem of the shortcoming of the dataset in deep neural networks by loading pre-trained BERT, fine-tuning it using the Persian rumor dataset, and combining with speech act and WS-based features. The empirical results of applying the model on Twitter and Telegram datasets demonstrated that BERT-SAWS can enhance the performance of the classifier from 2% to 18%. It indicates that speech act and WS alongside semantic contextual vectors are helpful features in the rumor verification task.

A Society’s Resilience to Conflict: An Analysis of Its Key Determinants in Conflict-Prone Rakhine State of Myanmar

This study explored contributors to a society’s resilience to conflict in a multiethnic community in Rakhine State of Myanmar. It examined how a society’s resilience to conflict was related to economic and sociocultural interactions between diverse communities, presence of security forces in a community’s neighboring areas, physical isolation from outsiders, people’s rumour verification practices, presence of civil society organizations (CSOs) and community based organizations (CBOs), and efficient community leadership. A survey of 1,668 respondents in 27 conflict-unaffected and conflict-affected villages and interview with 1,200 respondents was conducted among members of diverse ethnic groups across Rakhine State. Results suggested that there was a significant and positive correlation between a society’s resilience to conflict and each of ‘economic interactions’ (P = .000), ‘sociocultural interactions’ (P = .000), ‘presence of security forces in a community’s neighboring areas’ (P = .000), and ‘efficient community leadership’ (P = .000). Despite each individual of these independent variables being weakly correlated with the dependent variable, their combined effect strongly correlated with the dependent one. On the other hand, a society’s resilience to conflict negatively correlated with each of ‘physical isolation from outsiders’ (P = .001) and ‘people’s rumour verification practices' (P = .000). However, a society’s resilience to conflict hardly correlated with ‘presence of civil society organizations (CSOs) and community based organizations (CBOs)’, with only less than 1% of respondents in both conflict resilient and vulnerable areas articulating the significant role of these organizations in preventing intercommunal conflict in their areas.

RV-ML: An Effective Rumor Verification Scheme Based on Multi-Task Learning Model

Social platforms are full of rumors (i.e., unverified contents). Naturally, it is imperative but challenging to effectively determine the veracity of these rumors on popular social platforms. Previously deep learning based rumor verification schemes usually treat the issue as an independent and single task. Considering the rumor verification and stance classification are relevant tasks, we propose an effective Rumor verification scheme based on Multi-task learning Model, RV-ML, in which the shared long-short term memory (LSTM) layer for both rumor verification and stance classification can effectively deal with the sequential information for the original input, and generate macro-level virtual features, and the convolution neural network (CNN) layer uniquely designed for rumor verification task is used to mine local features from shared LSTM layer. Comparisons between our RV-ML and several typical rumor verification schemes on the real RumourEval and PHEME datasets demonstrate that our proposed scheme gains better performance for the task of rumor verification.

When Do People Verify and Share Health Rumors on Social Media? The Effects of Message Importance, Health Anxiety, and Health Literacy

This study explores the roles of perceived message importance, health anxiety, and health literacy in the relationship between message factors (message label and message valence) and behavioral intentions for rumor verification and sharing. 660 Twitter users responded to unverified information regarding the influenza vaccine. A 3 (label: none vs. news vs. rumor) × 2 (valence: positive vs. negative) online semi-experiment, with a survey to measure health anxiety and health literacy, showed the following results: First, perceived message importance mediated the relationship between message factors and behavioral intentions: only in the condition of the negative message, participants considered a news-labeled message more important than a rumor–labeled or a no-label message. Perceived message importance was associated with intentions to verify and share the message. Second, health anxiety interacted with perceived message importance only when predicting an intention to share the message. Last, healthy literacy interacted with perceived message importance when predicting intentions to both verify and share the message. The results will provide implications for health communication research and practices, especially on managing and controlling rumor dissemination on social media.

Rumour verification through recurring information and an inner-attention mechanism

Verification of online rumours is becoming an increasingly important task with the prevalence of event discussions on social media platforms. This paper proposes an inner-attention-based neural network model that uses frequent, recurring terms from past rumours to classify a newly emerging rumour as true, false or unverified. Unlike other methods proposed in related work, our model uses the source rumour alone without any additional information, such as user replies to the rumour or additional feature engineering. Our method outperforms the current state-of-the-art methods on benchmark datasets (RumourEval2017) by 3% accuracy and 6% F-1 leading to 60.7% accuracy and 61.6% F-1. We also compare our attention-based method to two similar models which however do not make use of recurrent terms. The attention-based method guided by frequent recurring terms outperforms this baseline on the same dataset, indicating that the recurring terms injected by the attention mechanism have high positive impact on distinguishing between true and false rumours. Furthermore, we perform out-of-domain evaluations and show that our model is indeed highly competitive compared to the baselines on a newly released RumourEval2019 dataset and also achieves the best performance on classifying fake and legitimate news headlines.