Abstract:The truth is significantly hampered by massive rumors that spread along with breaking news or popular topics. Since there is sufficient corpus gathered from the same domain for model training, existing rumor detection algorithms show promising performance on yesterday's news. However, due to a lack of substantial training data and prior expert knowledge, they are poor at spotting rumors concerning unforeseen events, especially those propagated in different languages (i.e., low-resource regimes). In this paper, we propose a unified contrastive transfer framework to detect rumors by adapting the features learned from well-resourced rumor data to that of the low-resourced with only few-shot annotations. More specifically, we first represent rumor circulated on social media as an undirected topology for enhancing the interaction of user opinions, and then train a Multi-scale Graph Convolutional Network via a unified contrastive paradigm to mine effective clues simultaneously from post semantics and propagation structure. Our model explicitly breaks the barriers of the domain and/or language issues, via language alignment and a novel domain-adaptive contrastive learning mechanism. To well-generalize the representation learning using a small set of annotated target events, we reveal that rumor-indicative signal is closely correlated with the uniformity of the distribution of these events. We design a target-wise contrastive training mechanism with three event-level data augmentation strategies, capable of unifying the representations by distinguishing target events. Extensive experiments conducted on four low-resource datasets collected from real-world microblog platforms demonstrate that our framework achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.

RumorMixer: Exploring Echo Chamber Effect and Platform Heterogeneity for Rumor Detection

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

Interpretable Graph Neural Network for Social Media Rumor Detection

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

Rumor Detection with Diverse Counterfactual Evidence

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

Rumor Detection Based on SAGNN: Simplified Aggregation Graph Neural Networks

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

Rumor Detection with a novel graph neural network approach

Rumor detection model with weighted GraphSAGE focusing on node location

Rumor Mitigation in Social Media Platforms with Deep Reinforcement Learning

A Rumor Detection Method Based on Multimodal Feature Fusion by a Joining Aggregation Structure

An End-to-End Rumor Detection Model Based on Feature Aggregation

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

Region-enhanced Deep Graph Convolutional Networks for Rumor Detection

Rumor2vec: A rumor detection framework with joint text and propagation structure representation learning

A Novel Fine-Grained Rumor Detection Algorithm with Attention Mechanism

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

Explainable Rumor Detection using Inter and Intra-feature Attention Networks

A Unified Contrastive Transfer Framework with Propagation Structure for Boosting Low-Resource Rumor Detection

Early Detection of Rumors in Heterogeneous Mobile Social Network.