Latent Multimodal Reconstruction for Misinformation Detection

2025-04-08

AI Summary

To address the challenges of scarce high-quality labeled data, insufficient realism and diversity in synthetic data generation, and underexploited potential of large vision-language models (LVLMs) in multimodal misinformation detection—particularly for image–text mislabeling—this paper proposes: (1) the first LVLM-driven paradigm for diverse mislabeled image generation, yielding the high-fidelity synthetic dataset “MisCaption This!”; (2) LAMAR, a latent-space cross-modal reconstruction framework that employs embedding-level reconstruction as strong auxiliary supervision, overcoming limitations of conventional classification-based paradigms; and (3) a multi-strategy architecture integrating mask, gate, and attention mechanisms, jointly optimized via end-to-end pretraining. Our approach establishes new state-of-the-art performance on NewsCLIPpings and VERITE, with significantly improved generalization to real-world scenarios. The code is publicly available.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Multimodal misinformation, such as miscaptioned images, where captions misrepresent an image's origin, context, or meaning, poses a growing challenge in the digital age. To support fact-checkers, researchers have been focusing on creating datasets and developing methods for multimodal misinformation detection (MMD). Due to the scarcity of large-scale annotated MMD datasets, recent studies leverage synthetic training data via out-of-context image-caption pairs or named entity manipulations; altering names, dates, and locations. However, these approaches often produce simplistic misinformation that fails to reflect real-world complexity, limiting the robustness of detection models trained on them. Meanwhile, despite recent advancements, Large Vision-Language Models (LVLMs) remain underutilized for generating diverse, realistic synthetic training data for MMD. To address this gap, we introduce"MisCaption This!", a training dataset comprising LVLM-generated miscaptioned images. Additionally, we introduce"Latent Multimodal Reconstruction"(LAMAR), a network trained to reconstruct the embeddings of truthful captions, providing a strong auxiliary signal to the detection process. To optimize LAMAR, we explore different training strategies (end-to-end training and large-scale pre-training) and integration approaches (direct, mask, gate, and attention). Extensive experiments show that models trained on"MisCaption This!"generalize better on real-world misinformation, while LAMAR sets new state-of-the-art on both NewsCLIPpings and VERITE benchmarks; highlighting the potential of LVLM-generated data and reconstruction-based approaches for advancing MMD. We release our code at: https://github.com/stevejpapad/miscaptioned-image-reconstruction

Problem

• Detecting multimodal misinformation in miscaptioned images
• Generating realistic synthetic training data for MMD
• Improving detection robustness with reconstruction-based approaches

Innovation

• LVLM-generated miscaptioned images dataset
• Latent Multimodal Reconstruction network
• Reconstruction of truthful caption embeddings