Abstract:The wide spread of rumors on social media has caused a negative impact on people's daily life, leading to potential panic, fear, and mental health problems for the public. How to debunk rumors as early as possible remains a challenging problem. Existing studies mainly leverage information propagation structure to detect rumors, while very few works focus on correlation among users that they may coordinate to spread rumors in order to gain large popularity. In this paper, we propose a new detection model, that jointly learns both the representations of user correlation and information propagation to detect rumors on social media. Specifically, we leverage graph neural networks to learn the representations of user correlation from a bipartite graph that describes the correlations between users and source tweets, and the representations of information propagation with a tree structure. Then we combine the learned representations from these two modules to classify the rumors. Since malicious users intend to subvert our model after deployment, we further develop a greedy attack scheme to analyze the cost of three adversarial attacks: graph attack, comment attack, and joint attack. Evaluation results on two public datasets illustrate that the proposed MODEL outperforms the state-of-the-art rumor detection models. We also demonstrate our method performs well for early rumor detection. Moreover, the proposed detection method is more robust to adversarial attacks compared to the best existing method. Importantly, we show that it requires a high cost for attackers to subvert user correlation pattern, demonstrating the importance of considering user correlation for rumor detection.

What problem does this paper attempt to address?

This paper attempts to address the issue of rumor propagation on social media. Specifically, the authors focus on how to identify and curb rumors at an early stage to reduce their impact on the daily lives of the public, such as potential panic, fear, and mental health issues. Existing research mainly relies on the structure of information dissemination to detect rumors but rarely considers the correlation between users who may collaboratively spread rumors to gain greater influence. Therefore, this paper proposes a new detection model that jointly learns user correlations and information dissemination representations to improve the accuracy of rumor detection on social media. ### Main Contributions: 1. **Proposed a new detection model**: This model detects rumors on social media by learning user correlations and information dissemination representations through Graph Neural Networks (GNN). 2. **Adversarial attack analysis**: A greedy attack scheme is used to generate fake edges and comments, exploring the cost of adversarial attacks. 3. **Performance evaluation**: Experiments were conducted on two public datasets, and the results show that the model outperforms existing methods in terms of detection accuracy and F1 score, and has higher robustness against adversarial attacks. ### Method Overview: - **User correlation representation**: A user-tweet bipartite graph is constructed, and a two-layer Graph Convolutional Network (GCN) is used to learn the representation of the source tweet. - **Information dissemination representation**: A tree-structured information dissemination model is used, and GCN is also employed to learn the information dissemination representation. - **Joint classification**: The user correlation representation and information dissemination representation are combined and input into a fully connected layer for classification, with the final classification result output through a softmax layer. ### Experimental Results: - Extensive experiments were conducted on the Twitter15 and Twitter16 datasets, and the results show that the model outperforms baseline methods in terms of detection accuracy and F1 score. - Adversarial attack analysis shows that attackers need higher costs to deceive the model, demonstrating its stronger robustness. ### Conclusion: The model proposed in this paper, which jointly learns user correlations and information dissemination representations, performs excellently in rumor detection on social media, especially in early detection. Future work will focus on further optimizing the representation learning of textual information to enhance model performance.

Rumor Detection with a novel graph neural network approach

Interpretable Graph Neural Network for Social Media Rumor Detection

Detecting the Rumor Patterns Integrating Features of User, Content, and the Spreading Structure.

Rumor detection on social media using hierarchically aggregated feature via graph neural networks

Epidemiology-informed Network for Robust Rumor Detection

Graph Convolution-Based Joint Learning of Rumor with Content, User Credibility, Propagation Context, and Cognitive as Well as Emotion Signals

Predicting Viral Rumors and Vulnerable Users for Infodemic Surveillance

Early detection of rumors based on source tweet-word graph attention networks.

A Novel Fine-Grained Rumor Detection Algorithm with Attention Mechanism

Rumor Detection on Social Media with Temporal Propagation Structure Optimization

Rumor Detection on Social Media with Bi-Directional Graph Convolutional Networks

Towards Robust Rumor Detection with Graph Contrastive and Curriculum Learning

Modeling the Uncertainty of Information Propagation for Rumor Detection: A Neuro-Fuzzy Approach

Joint learning of structural and textual information on propagation network by graph attention networks for rumor detection

Rumor Detection with Diverse Counterfactual Evidence

Rumor Detection Based on Knowledge Enhancement and Graph Attention Network

Heterogeneous Graph Attention Networks with Bi-directional Information Propagation for Rumor Detection

Region-enhanced Deep Graph Convolutional Networks for Rumor Detection

DDGCN: Dual Dynamic Graph Convolutional Networks for Rumor Detection on Social Media

A Novel and High-Accuracy Rumor Detection Approach Using Kernel Subtree and Deep Learning Networks

Graph-aware multi-feature interacting network for explainable rumor detection on social network