Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EMIF: Evidence-aware Multi-source Information Fusion Network for Explainable Fake News Detection

Qingxing Dong, Mengyi Zhang, Shiyuan Wu, Xiaozhen Wu

TL;DR

This paper tackles the problem of robust, explainable fake news detection by integrating multiple evidence sources. The proposed EMIF model jointly leverages user comments and relevant news, using a co-attention mechanism to capture semantic conflicts and a divergence-based selection to assemble diverse, objective evidence; an inconsistency loss ties the two streams together during fusion. Empirical results on the FibVID COVID-19 dataset show that EMIF outperforms strong baselines and remains robust when one information source is unavailable, with ablation analyses highlighting the critical roles of comments, relevant news, and the co-attention framework. The approach advances practical explainability by visualizing attention-driven evidence and offers a scalable, multi-source paradigm for trustworthy fake news verification with potential extensions to additional modalities.

Abstract

Extensive research on automatic fake news detection has been conducted due to the significant detrimental effects of fake news proliferation. Most existing approaches rely on a single source of evidence, such as comments or relevant news, to derive explanatory evidence for decision-making, demonstrating exceptional performance. However, their single evidence source suffers from two critical drawbacks: (i) noise abundance, and (ii) resilience deficiency. Inspired by the natural process of fake news identification, we propose an Evidence-aware Multi-source Information Fusion (EMIF) network that jointly leverages user comments and relevant news to make precise decision and excavate reliable evidence. To accomplish this, we initially construct a co-attention network to capture general semantic conflicts between comments and original news. Meanwhile, a divergence selection module is employed to identify the top-K relevant news articles with content that deviates the most from the original news, which ensures the acquisition of multiple evidence with higher objectivity. Finally, we utilize an inconsistency loss function within the evidence fusion layer to strengthen the consistency of two types of evidence, both negating the authenticity of the same news. Extensive experiments and ablation studies on real-world dataset FibVID show the effectiveness of our proposed model. Notably, EMIF shows remarkable robustness even in scenarios where a particular source of information is inadequate.

EMIF: Evidence-aware Multi-source Information Fusion Network for Explainable Fake News Detection

TL;DR

This paper tackles the problem of robust, explainable fake news detection by integrating multiple evidence sources. The proposed EMIF model jointly leverages user comments and relevant news, using a co-attention mechanism to capture semantic conflicts and a divergence-based selection to assemble diverse, objective evidence; an inconsistency loss ties the two streams together during fusion. Empirical results on the FibVID COVID-19 dataset show that EMIF outperforms strong baselines and remains robust when one information source is unavailable, with ablation analyses highlighting the critical roles of comments, relevant news, and the co-attention framework. The approach advances practical explainability by visualizing attention-driven evidence and offers a scalable, multi-source paradigm for trustworthy fake news verification with potential extensions to additional modalities.

Abstract

Extensive research on automatic fake news detection has been conducted due to the significant detrimental effects of fake news proliferation. Most existing approaches rely on a single source of evidence, such as comments or relevant news, to derive explanatory evidence for decision-making, demonstrating exceptional performance. However, their single evidence source suffers from two critical drawbacks: (i) noise abundance, and (ii) resilience deficiency. Inspired by the natural process of fake news identification, we propose an Evidence-aware Multi-source Information Fusion (EMIF) network that jointly leverages user comments and relevant news to make precise decision and excavate reliable evidence. To accomplish this, we initially construct a co-attention network to capture general semantic conflicts between comments and original news. Meanwhile, a divergence selection module is employed to identify the top-K relevant news articles with content that deviates the most from the original news, which ensures the acquisition of multiple evidence with higher objectivity. Finally, we utilize an inconsistency loss function within the evidence fusion layer to strengthen the consistency of two types of evidence, both negating the authenticity of the same news. Extensive experiments and ablation studies on real-world dataset FibVID show the effectiveness of our proposed model. Notably, EMIF shows remarkable robustness even in scenarios where a particular source of information is inadequate.
Paper Structure (18 sections, 13 equations, 4 figures, 3 tables)

This paper contains 18 sections, 13 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Models for fake news detection
  4. Information sources in fake news detection
  5. Methodology
  6. Background & notations
  7. Model Architecture
  8. Input encoding layer
  9. Co-attention mechanism
  10. Divergence selection
  11. Evidence fusion layer
  12. Experiments
  13. Datasets
  14. Baselines
  15. Overall performance
  16. ...and 3 more sections

Figures (4)

  • Figure 1: A piece of news and its related comments on social media. The green and red comments express supporting and opposing viewpoints, respectively. But only Comment-2 and Comment-4 point out the core false part of this fake news.
  • Figure 2: The architecture of EMIF model.
  • Figure 3: Impact analysis of each component of EMIF for fake news detection
  • Figure 4: Explainability via visualization of attention weights in EMIF. The label [True/False] indicates the verdict of original news. (a) are user comments (the first 3 are explainable comments captured can be used as evidence), (b) are original news (labeled false, darker shades correspond to higher-weighted words) and selected relevant news (only 4 items were presented).