Abstract:
The rapid proliferation of disinformation on social media platforms poses significant challenges to public trust, societal stability, and informed decisionmaking. Traditional fake news detection approaches often rely on textual content alone, which limits their effectiveness against increasingly sophisticated and multimodal disinformation campaigns. In this paper, we propose a multimodal framework that integrates textual, visual, and contextual features to enhance the accuracy and robustness of fake news identification. Our approach leverages advanced natural language processing (NLP) models for semantic analysis, computer vision techniques for image verification, and graph-based methods for capturing user interaction patterns. By fusing these heterogeneous modalities, the system provides a comprehensive assessment of news veracity in real time. Experimental results on benchmark social media datasets demonstrate that the proposed multimodal model significantly outperforms unimodal baselines, achieving improved precision, recall, and F1- scores. This research contributes to building more resilient systems for combating disinformation, with potential applications in fact-checking, digital journalism, and online content moderation.
Page(s):
26-26
DOI:
DOI not available
Published:
Journal: 4th International Conference of Sciences “Revamped Scientific Outlook of 21st Century, 2025” , November 12,2025, Volume: 1, Issue: 1, Year: 2025
Keywords:
Computer vision
,
Natural language processing NLP
,
Semantic Analysis
,
Multimodal Disinformation