Jointly Embedding the Local and Global Relations of Heterogeneous Graph for Rumor Detection

Abstract

The development of social media has revolutionized the way people communicate, share information and make decisions, but it also provides an ideal platform for publishing and spreading rumors. Existing rumor detection methods focus on finding clues from text content, user profiles, and propagation patterns. However, the local semantic relation and global structural information in the message propagation graph have not been well utilized by previous works.In this paper, we present a novel global-local attention network (GLAN) for rumor detection, which jointly encodes the local semantic and global structural information. We first generate a better integrated representation for each source tweet by fusing the semantic information of related retweets with the attention mechanism. Then, we model the global relationships among all source tweets, retweets, and users as a heterogeneous graph to capture the rich structural information for rumor detection. We conduct experiments on three real-world datasets, and the results demonstrate that GLAN significantly outperforms the state-ofthe-art models in both rumor detection and early detection scenarios.

Deep spatial–temporal structure learning for rumor detection on Twitter

“…Note that although there emerge some strong methods recently, such as PPC [16] and GLAN [17], they use user information to guide model learning. Since we focus on, to what extent, the rumor detection can be solved if only the spatial-temporal structure information is available, we do not compare with them in our experiments.…”

“…We conduct experiments on two publicly available Twitter datasets: Twitter15 and Twitter16, which have been widely adopted as standard data in the field of rumor detection Ma et al [11], Liu and Wu [16], Ma et al [3], Yuan et al [17]. Twitter15 dataset contains 1490 tweets propagations and Twitter16 contains 818 tweets propagations with their more details shown in Table 1.…”

Rumor Detection Based on SAGNN: Simplified Aggregation Graph Neural Networks

“…As suggested in the introduction, rumor detection related work can be divided into following categories [5]: 1 See [2], [5] for a more detailed review of other feature-based and propagation tree related methods.…”

“…A global-local attention network (GLAN) was proposed in [5] for rumor detection, which jointly encodes the local semantic and global structural information. A bidirectional graph convolutional network architecture was proposed in [6] to further enhance the rumor detection performance.…”

“…Latest efforts focus on applying graph learning techniques for rumor detection [4], [5], [6], due to the development of graph neural networks and the fact that posts on social media are naturally structured as graphs.…”