Rumor Detection on Social Media with Graph Adversarial Contrastive Learning

Sun, Tiening; Qian, Zhong; Dong, Sujun; Li, Peifeng; Zhu, Qiaoming

doi:10.1145/3485447.3511999

Cited by 59 publications

(26 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to classification loss, some models (eg. RDEA [27], GACL [28]) utilises additional loss function to enhance the detection, which will be declared later.…”

Section: Propagation-based Methodsmentioning

confidence: 99%

“…For example, Zhang et al [67] use Graph Attention Networks (GAT) [34] to model the propagation graph; others such as [68], [69] use graph auto-encoders and leverage their strong learning capabilities. GACL [28] uses an adversarial contrastive learning method (discussed in Section V-A2), which also follows the bi-directional graph modelling framework.…”

Section: A Static Graph-based Methodsmentioning

confidence: 99%

“…RDEA [27] and GACL [28] follow Ma et al [70] and further introduce contrastive learning into graph-based fake news detection. They train more robust propagation-based models with contrastive learning.…”

Section: A Static Graph-based Methodsmentioning

confidence: 99%

“…Enhance edge connections by Gaussian distributions and KL-divergence. EDEA [27], GACL [28], RDCL [29], CCFD [30] claim, comments responsive propagation (with tuning)…”

Section: Gcnmentioning

confidence: 99%

See 3 more Smart Citations

Heterogeneous graph neural networks with denoising for graph embeddings

Dong

Zhang

Pang

et al. 2022

Knowledge-Based Systems

View full text Add to dashboard Cite

“…In addition to classification loss, some models (eg. RDEA [27], GACL [28]) utilises additional loss function to enhance the detection, which will be declared later.…”

Section: Propagation-based Methodsmentioning

confidence: 99%

Section: A Static Graph-based Methodsmentioning

confidence: 99%

Section: A Static Graph-based Methodsmentioning

confidence: 99%

“…Enhance edge connections by Gaussian distributions and KL-divergence. EDEA [27], GACL [28], RDCL [29], CCFD [30] claim, comments responsive propagation (with tuning)…”

Section: Gcnmentioning

confidence: 99%

See 2 more Smart Citations

Heterogeneous graph neural networks with denoising for graph embeddings

Dong

Zhang

Pang

et al. 2022

Knowledge-Based Systems

View full text Add to dashboard Cite

“…GACL (Sun et al, 2022): A SOTA in-topic false information detection model with supervised contrastive and adversarial learning. The method utilizes supervised contrastive learning to improve the model's generalization and introduces adversarial learning to boost the robustness of the model.…”

Section: Baseline and Sotasmentioning

confidence: 99%

Towards Robust False Information Detection on Social Networks with Contrastive Learning

et al. 2022

Proceedings of the 31st ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

Current works about false information detection based on conversation graphs on social networks focus primarily on two research streams from the standpoint of topic distribution: intopic and cross-topic techniques, which assume that the data topic distribution is identical or cross, respectively. This signifies that all test data topics are seen or unseen by the model. However, these assumptions are too harsh for actual social networks that contain both seen and unseen topics simultaneously, hence restricting their practical application. In light of this, this paper develops a novel open-topic scenario that is better suited to actual social networks. In this open-topic scenario, we empirically find that the existing models suffer from impairment in the detection performance for seen or unseen topic data, resulting in poor overall model performance. To address this issue, we propose a novel Contrastive Adversarial Learning Network, CALN, that employs an unsupervised topic clustering method to capture topic-specific features to enhance the model's performance for seen topics and an unsupervised adversarial learning method to align data representation distributions to enhance the model's generalisation to unseen topics. Experiments on two benchmark datasets and a variety of graph neural networks demonstrate the effectiveness of our approach.

show abstract