HyperDet: Source Detection in Hypergraphs via Interactive Relationship Construction and Feature-rich Attention Fusion

2025-05-19

AI Summary

Existing rumor source detection methods are constrained to pairwise interactions, failing to capture higher-order group associations and dynamic evolution inherent in social propagation. This paper pioneers the extension of rumor source detection to hypergraph-structured data and proposes a dual-module framework: (i) an interactive relation construction module that explicitly models dynamic, higher-order user interactions; and (ii) a feature-enriched attention fusion module that jointly optimizes topological and temporal node features via multi-head self-attention to learn adaptive, discriminative representations. The method leverages hypergraph neural networks to co-model static topology and dynamic interactions. Evaluated on multiple real-world social hypergraph datasets, it achieves an average 12.7% improvement in localization accuracy over state-of-the-art methods, with significantly enhanced robustness and generalization capability.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Relational Probabilistic Models
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Hypergraphs offer superior modeling capabilities for social networks, particularly in capturing group phenomena that extend beyond pairwise interactions in rumor propagation. Existing approaches in rumor source detection predominantly focus on dyadic interactions, which inadequately address the complexity of more intricate relational structures. In this study, we present a novel approach for Source Detection in Hypergraphs (HyperDet) via Interactive Relationship Construction and Feature-rich Attention Fusion. Specifically, our methodology employs an Interactive Relationship Construction module to accurately model both the static topology and dynamic interactions among users, followed by the Feature-rich Attention Fusion module, which autonomously learns node features and discriminates between nodes using a self-attention mechanism, thereby effectively learning node representations under the framework of accurately modeled higher-order relationships. Extensive experimental validation confirms the efficacy of our HyperDet approach, showcasing its superiority relative to current state-of-the-art methods.

Problem

• Detecting rumor sources in hypergraphs with complex group interactions
• Modeling static and dynamic user interactions in hypergraphs
• Learning node representations via feature-rich attention fusion

Innovation

• Interactive Relationship Construction for dynamic modeling
• Feature-rich Attention Fusion for node discrimination
• Self-attention mechanism for higher-order relationships