Vision-language method improves fake news detection with reusable concepts
ReVR: Dual-Path Concept Reasoning for Multimodal Fake News Detection
Multimedia
Summary
Detecting fake news that uses both pictures and text is difficult because the system needs to understand both types of information together. The authors created a method called ReVR that learns and uses clear concepts to decide if news is true or not. ReVR uses two different ways to check these concepts: one checks all known ideas broadly, and the other focuses closely on specific doubts. When these ways disagree, a resolver picks the best answer. Tests show ReVR is good at catching fake news and works well on new types it hasn’t seen before.
What this means in practice
- •For media monitoring teams: Use reusable verification concepts to enhance detection of multimodal fake news across various news sources.
- •For social media platform engineers: Integrate dual-path concept reasoning to improve automated identification of fake posts combining images and text.
Authors
Zhikai Tan, Yuzhou Yang, Qichao Ying, Pinjie Xu, Sheng Li, Zhenxing Qian, Xinpeng Zhang
Abstract
Vision-language models (VLMs) support multimodal fake news detection (FND) by producing explicit analyses. Recent methods further improve interpretability by organizing verification knowledge into explicit concepts. However, two questions remain: how to improve the reliability and applicability of verification concepts, and how to effectively apply reusable concepts to verify unseen news. We propose \textbf{ReVR}, a dual-path reasoning framework that constructs and applies reusable verification concepts for multimodal fake news detection. An agentic workflow grounds and consolidates candidate concepts, while statistical profiles characterize their historical behavior. During inference, a coverage-oriented path aggregates evidence from the complete concept library using a trainable encoder, while a query-focused path prompts a frozen VLM to reason over selected concepts and their observations. A learned conflict resolver selects between the two predictions when they disagree. Experiments on fake news benchmarks demonstrate the effectiveness of the method regarding detection performance and generalizability.