Abstract:The rapid spread of rumors on social media has posed significant challenges to maintaining public trust and information integrity. Since an information cascade process is essentially a propagation tree, recent rumor detection models leverage graph neural networks to additionally capture information propagation patterns, thus outperforming text-only solutions. Given the variations in topics and social impact of the root node, different source information naturally has distinct outreach capabilities, resulting in different heights of propagation trees. This variation, however, impedes the data-driven design of existing graph-based rumor detectors. Given a shallow propagation tree with limited interactions, it is unlikely for graph-based approaches to capture sufficient cascading patterns, questioning their ability to handle less popular news or early detection needs. In contrast, a deep propagation tree is prone to noisy user responses, and this can in turn obfuscate the predictions. In this paper, we propose a novel Epidemiology-informed Network (EIN) that integrates epidemiological knowledge to enhance performance by overcoming data-driven methods sensitivity to data quality. Meanwhile, to adapt epidemiology theory to rumor detection, it is expected that each users stance toward the source information will be annotated. To bypass the costly and time-consuming human labeling process, we take advantage of large language models to generate stance labels, facilitating optimization objectives for learning epidemiology-informed representations. Our experimental results demonstrate that the proposed EIN not only outperforms state-of-the-art methods on real-world datasets but also exhibits enhanced robustness across varying tree depths.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to address the challenges brought by the rapid spread of rumors on social media platforms, especially how to improve the accuracy and robustness of rumor detection under different depths of the spreading tree. Specifically, the paper focuses on the following key issues: 1. **Data quality issues in rumor detection**: - Existing rumor - detection models based on Graph Neural Networks (GNN) are very sensitive to data quality. When the spreading tree is shallow, due to limited interactions, these models have difficulty capturing sufficient spreading patterns; when the spreading tree is deep, the noise of user responses will increase, thus affecting the accuracy of prediction. 2. **Performance differences under different spreading - tree depths**: - Source information with different topics and social influence will lead to different heights of the spreading tree, which in turn affects the effectiveness of rumor detection. Existing methods perform inconsistently when dealing with shallow and deep spreading trees and lack robustness across different tree depths. 3. **Challenges brought by the structural complexity of the spreading tree**: - The structural complexity of the spreading tree makes it difficult for existing methods to simultaneously meet the needs of early detection and handle complex spreading patterns. For example, in the information released by new or less - influential users, the spreading tree is usually shallow, while the content of social celebrities may form a deeper spreading tree. To solve these problems, the author proposes a new Epidemiology - informed Network (EIN), which enhances the performance of rumor detection by integrating epidemiological theory. The main innovations of EIN include: - **Introducing the epidemiological model (eUSD)**: Model the rumor - spreading process as the transformation of three states - unknown, support, and denial in the environment, in order to more accurately capture the dynamics of rumor spreading. - **Using large - language models (LLM) to generate stance labels**: Generate stance labels for each post through LLM to optimize the learning of epidemiologically - informed representations without manual annotation. - **Combining graph neural networks and epidemiological models**: Seamlessly integrate the EIN framework with existing graph - neural - network rumor detectors, improving the robustness and accuracy of the model. Through these improvements, EIN not only performs well on multiple real - world datasets but also shows stronger robustness under different spreading - tree depths.

Epidemiology-informed Network for Robust Rumor Detection

Interpretable Graph Neural Network for Social Media Rumor Detection

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

Rumor Detection with a novel graph neural network approach

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

Predicting Viral Rumors and Vulnerable Users for Infodemic Surveillance

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

Rumor Detection on Social Media with Temporal Propagation Structure Optimization

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

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

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

Continuous-Time Dynamic Graph Networks Integrated with Knowledge Propagation for Social Media Rumor Detection

RP-DNN: A Tweet level propagation context based deep neural networks for early rumor detection in Social Media

Towards Robust Rumor Detection with Graph Contrastive and Curriculum Learning

A Rumor Detection Model Incorporating Propagation Path Contextual Semantics and User Information

Rumor Detection Based on Knowledge Enhancement and Graph Attention Network

Region-enhanced Deep Graph Convolutional Networks for Rumor Detection

Neural network approaches for rumor stance detection: Simulating complex rumor propagation systems

NIT: Searching for rumors in social network through Neighborhood Information Transmission

Rumor Detection Based on Knowledge Graph and Graph Attention Network

Rooting Our Rumor Sources in Online Social Networks: the Value of Diversity from Multiple Observations