The rise of multimodal fake news threatens reliable information dissemination by exploiting multiple modalities to create deceptive, engaging content, significantly impacting society safety. Existing methods still face challenges in cross-modal alignment (e.g., semantic inconsistencies, complex visual-semantic relations) and are vulnerable to low-quality or noisy samples. To address these, we propose Cross-Modal Alignment with Visual Reasoning Prompting (CMA-VRP) for multimodal fake news detection. Specifically, we model text and image entities with graphs to capture fine-grained semantic interactions and enhance cross-modal consistency through graph contrastive learning. Unlike methods relying on shallow image features (e.g., edges, textures), we leverage large language models (LLMs) and large vision-language models (LVLMs) to capture deep visual-semantic attributes related to reasoning (e.g., actions, scenes). Based on graph modeling and visual reasoning features, we perform graph-based cross-modal semantic fusion to unify textual and visual representations and cross-modal cycle alignment to align modality distributions by reducing semantic discrepancies, filtering modality-specific noise, and extracting invariant representations across domains. These steps enable the model to obtain semantically consistent and modality-invariant features. Extensive experiments demonstrate that our model outperforms existing methods in multimodal fake news detection and shows strong robustness against noisy samples.

In multimodal fact-checking, advanced large multimodal models (LMMs) struggle to capture and integrate the complex relationships between text and images. A potential solution is to generate reasoning support text to optimize reasoning and integrate evidence. However, existing generation approaches rely heavily on high-quality data annotations for training, which are costly and limited in scalability, hindering responsiveness to evolving misinformation. To address these issues, we propose CoReS, a novel zero-shot multimodal fact-checking model based on Conceptual Reasoning Support-leveraging key concepts from evidence to guide the reasoning process and improve decision-making. This model includes a reasoning support text generation module that extracts key concepts (critical elements that significantly impact the judgment outcome) from raw textual evidence via retrieval and filtering. By using a Conceptual Reasoning LM, CoReS generates reasoning support texts framed around core key concepts that are semantically consistent with multimodal evidence, linking key clues, thus replacing redundant and complex evidence for fact-checking. The reasoning support texts generated by CoReS effectively distill complex evidence relationships and integrate important reasoning information, allowing the judgment model to provide clear and accurate judgments. Evaluations on benchmark datasets and the new multi-domain MultiVerify dataset demonstrate that CoReS excels in accuracy, generalization, and scalability.

The advent of the new media era has greatly changed the original pattern of science communication, and social networks have become a new front to promote the Public Understanding of Science. The integration of social networks and scientific communication has dissolved the central focus of scientific communication, moving toward a more diversified subject centered around the public networks. Due to the special nature of science communication, the key nodes in the community network often play a crucial role. Specifically, by identifying key nodes and analyzing the trend of grouping them into clusters, the stability and robustness of the complex network of science communication can be significantly improved. This is of great significance for expanding the influence of science communication and attracting relevant practitioners to join. However, the sensitive attributes and potential influence factors of community network nodes can lead to deviations in key node mining using existing algorithms. From the perspective of algorithm fairness, we combine graph neural networks and counterfactual theory to address this issue in this work and propose a framework for key node mining for the characteristics of science communication,CounteRfActual Fairness discriminaTor (CRAFT). Experimental results on real data demonstrate the prediction and fairness results and the efficacy of CRAFT in learning causal representations of latent factors.