Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Out-of-distribution Rumor Detection via Test-Time Adaptation

Xiang Tao, Mingqing Zhang, Qiang Liu, Shu Wu, Liang Wang

TL;DR

The paper tackles the problem of rumor detection under distribution shifts across topics, platforms, and languages by introducing TARD, a test-time adaptation framework that operates on news propagation graphs. TARD fuses supervised rumor classification with a graph-based self-supervised learning task and performs test-time updates to the shared graph encoder and SSL head, guided by an adaptive constraint to prevent representation distortion. The key contributions are (i) the formulation of an OOD rumor detection framework on propagation graphs, (ii) the integration of test-time adaptation with graph contrastive learning, and (iii) an adaptive constraint that stabilizes embeddings during test-time training. Empirical results on two real-world, cross-domain rumor datasets demonstrate that TARD achieves state-of-the-art performance under distribution shifts, highlighting its practical potential for robust, real-world rumor detection across diverse social media environments.

Abstract

Due to the rapid spread of rumors on social media, rumor detection has become an extremely important challenge. Existing methods for rumor detection have achieved good performance, as they have collected enough corpus from the same data distribution for model training. However, significant distribution shifts between the training data and real-world test data occur due to differences in news topics, social media platforms, languages and the variance in propagation scale caused by news popularity. This leads to a substantial decline in the performance of these existing methods in Out-Of-Distribution (OOD) situations. To address this problem, we propose a simple and efficient method named Test-time Adaptation for Rumor Detection under distribution shifts (TARD). This method models the propagation of news in the form of a propagation graph, and builds propagation graph test-time adaptation framework, enhancing the model's adaptability and robustness when facing OOD problems. Extensive experiments conducted on two group datasets collected from real-world social platforms demonstrate that our framework outperforms the state-of-the-art methods in performance.

Out-of-distribution Rumor Detection via Test-Time Adaptation

TL;DR

The paper tackles the problem of rumor detection under distribution shifts across topics, platforms, and languages by introducing TARD, a test-time adaptation framework that operates on news propagation graphs. TARD fuses supervised rumor classification with a graph-based self-supervised learning task and performs test-time updates to the shared graph encoder and SSL head, guided by an adaptive constraint to prevent representation distortion. The key contributions are (i) the formulation of an OOD rumor detection framework on propagation graphs, (ii) the integration of test-time adaptation with graph contrastive learning, and (iii) an adaptive constraint that stabilizes embeddings during test-time training. Empirical results on two real-world, cross-domain rumor datasets demonstrate that TARD achieves state-of-the-art performance under distribution shifts, highlighting its practical potential for robust, real-world rumor detection across diverse social media environments.

Abstract

Due to the rapid spread of rumors on social media, rumor detection has become an extremely important challenge. Existing methods for rumor detection have achieved good performance, as they have collected enough corpus from the same data distribution for model training. However, significant distribution shifts between the training data and real-world test data occur due to differences in news topics, social media platforms, languages and the variance in propagation scale caused by news popularity. This leads to a substantial decline in the performance of these existing methods in Out-Of-Distribution (OOD) situations. To address this problem, we propose a simple and efficient method named Test-time Adaptation for Rumor Detection under distribution shifts (TARD). This method models the propagation of news in the form of a propagation graph, and builds propagation graph test-time adaptation framework, enhancing the model's adaptability and robustness when facing OOD problems. Extensive experiments conducted on two group datasets collected from real-world social platforms demonstrate that our framework outperforms the state-of-the-art methods in performance.
Paper Structure (20 sections, 6 equations, 3 figures, 1 table)

This paper contains 20 sections, 6 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. The proposed TARD Model
  3. Problem formulation
  4. Framework
  5. Self-supervised Learning Task
  6. Train Phase
  7. Test-Time Training Phase
  8. Adaptation Constraint
  9. Test Phase
  10. Experiments
  11. Experimental Setups
  12. Datasets
  13. Baselines
  14. Experimental Settings
  15. Overall Performance
  16. ...and 5 more sections

Figures (3)

  • Figure 1: The overall framework of our proposed TARD model.During the training phase, all model parameters are updated. In the test-time training phase, only the parameters of the shared feature extractor and the SSL head are updated, allowing the model to adapt to the data distribution of the test dataset. During the testing phase, the updated shared feature extractor and classification head are used to classify the test dataset.
  • Figure 2: Ablation studies on our proposed model
  • Figure 3: Sensitivity analysis of hyperparameters $\alpha_1$ and $\alpha_2$.