Two Birds with One Stone: Improving Rumor Detection by Addressing the Unfairness Issue

2024-12-30

AI Summary

To address group-wise performance disparities and fairness deficits in rumor detection caused by sensitive attributes (e.g., gender, region), this paper proposes a two-stage invariant learning framework that operates without labeled sensitive attributes. First, it identifies confounding sensitive attributes via unsupervised discovery; then, it jointly applies invariant risk minimization (IRM) and multi-group representation disentanglement to learn robust, group-agnostic representations that preserve task-relevant information. The method innovatively integrates deconfounding and invariant learning into rumor detection, augmented by a lightweight adapter module and causal inference modeling. Evaluated on multiple benchmark datasets, it achieves average accuracy gains of 2.1–4.7%, reduces equal opportunity difference by up to 63%, and demonstrates strong plug-and-play compatibility and cross-dataset generalizability. To our knowledge, this is the first approach to simultaneously enhance both predictive accuracy and group fairness in rumor detection.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

The degraded performance and group unfairness caused by confounding sensitive attributes in rumor detection remains relatively unexplored. To address this, we propose a two-step framework. Initially, it identifies confounding sensitive attributes that limit rumor detection performance and cause unfairness across groups. Subsequently, we aim to learn equally informative representations through invariant learning. Our method considers diverse sets of groups without sensitive attribute annotations. Experiments show our method easily integrates with existing rumor detectors, significantly improving both their detection performance and fairness.

Problem

• Rumor Detection
• Accuracy Improvement
• Fairness in Processing

Innovation

• Two-step Method
• Invariance Learning
• Fairness Enhancement