DiFaR: Enhancing Multimodal Misinformation Detection with Diverse, Factual, and Relevant Rationales

2025-08-14

AI Summary

Current large vision-language models (LVLMs) generate rationales for multimodal misinformation detection that suffer from three key limitations: insufficient diversity, poor factual consistency (i.e., severe hallucination), and weak relevance (e.g., inclusion of irrelevant or contradictory content). To address these issues, we propose DiFaR—a detector-agnostic framework that jointly optimizes rationale diversity, factuality, and relevance via five chain-of-thought prompting strategies and a lightweight sentence-level post-processing module. Our core innovation lies in a dual-dimension scoring mechanism—evaluating both factuality and relevance—to guide sentence selection, coupled with end-to-end co-optimization of prompting strategies and filtering. Experiments across four mainstream benchmarks demonstrate that DiFaR outperforms four categories of baselines by up to 5.9% in rationale quality and boosts the performance of existing detectors by up to 8.7%, significantly enhancing both rationale fidelity and detection robustness.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models

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
• Responsible Web: Mitigating misinformation and disinformation

Abstract

Generating textual rationales from large vision-language models (LVLMs) to support trainable multimodal misinformation detectors has emerged as a promising paradigm. However, its effectiveness is fundamentally limited by three core challenges: (i) insufficient diversity in generated rationales, (ii) factual inaccuracies due to hallucinations, and (iii) irrelevant or conflicting content that introduces noise. We introduce DiFaR, a detector-agnostic framework that produces diverse, factual, and relevant rationales to enhance misinformation detection. DiFaR employs five chain-of-thought prompts to elicit varied reasoning traces from LVLMs and incorporates a lightweight post-hoc filtering module to select rationale sentences based on sentence-level factuality and relevance scores. Extensive experiments on four popular benchmarks demonstrate that DiFaR outperforms four baseline categories by up to 5.9% and boosts existing detectors by as much as 8.7%. Both automatic metrics and human evaluations confirm that DiFaR significantly improves rationale quality across all three dimensions.

Problem

• Addressing insufficient diversity in generated rationales
• Mitigating factual inaccuracies from model hallucinations
• Reducing irrelevant or conflicting content introducing noise

Innovation

• Chain-of-thought prompts for diverse reasoning
• Lightweight filtering module for factuality
• Sentence-level scoring for relevance selection