The prevalence and perniciousness of fake news have been a critical issue on the Internet, which stimulates the development of automatic fake news detection in turn. In this paper, we focus on the evidence-based fake news detection, where several evidences are utilized to probe the veracity of news (i.e., a claim). Most previous methods first employ sequential models to embed the semantic information and then capture the claim-evidence interaction based on different attention mechanisms. Despite their effectiveness, they still suffer from three weaknesses. First, due to the inherent drawbacks of sequential models, they fail to integrate the relevant information that is scattered far apart in evidences for veracity checking. Second, they underestimate much redundant information contained in evidences that may be useless or even harmful. Third, insufficient data utilization limits the separability and reliability of representations captured by the model, which are sensitive to local evidence. To solve these problems, we propose a unified G raph-based s E mantic structure mining framework with Con TRA stive L earning, namely GETRAL in short. Specifically, different from the existing work that treats claims and evidences as sequences, we first model them as graph-structured data and capture the long-distance semantic dependency among dispersed relevant snippets via neighborhood propagation. After obtaining contextual semantic information, our model reduces information redundancy by performing graph structure learning. Then the fine-grained semantic representations are fed into the downstream claim-evidence interaction module for predictions. Finally, the supervised contrastive learning accompanied with adversarial augmented instances is applied to make full use of data and strengthen the representation learning. Comprehensive experiments have demonstrated the superiority of GETRAL over the state-of-the-arts and validated the efficacy of semantic mining with graph structure and contrastive learning.

Evidence-aware fake news detection aims to conduct reasoning between news and evidences, which are retrieved based on news content, to find uniformity or inconsistency. However, we find evidence-aware detection models suffer from biases, i.e., spurious correlations between news/evidence contents and true/fake news labels, and are hard to be generalized to Out-Of-Distribution (OOD) situations. To deal with this, we propose a novel Dual Adversarial Learning (DAL) approach. We incorporate news-aspect and evidence-aspect debiasing discriminators, whose targets are both true/fake news labels, in DAL. Then, DAL reversely optimizes news-aspect and evidence-aspect debiasing discriminators to mitigate the impact of news and evidence content biases. At the same time, DAL also optimizes the main fake news predictor, so that the news-evidence interaction module can be learned. This process allows us to teach evidence-aware fake news detection models to better conduct news-evidence reasoning, and minimize the impact of content biases. To be noted, our proposed DAL approach is a plug-and-play module that works well with existing backbones. We conduct comprehensive experiments under two OOD settings, and plug DAL in four evidence-aware fake news detection backbones. Results demonstrate that, DAL significantly and stably outperforms the original backbones and some competitive debiasing methods.

With the rapid expansion of multimodal content in online social media, automatic detection of multimodal fake news has received much attention. Multimodal joint training commonly used in existing methods is expected to benefit from thoroughly leveraging cross-modal features, yet these methods still suffer from insufficient learning of uni-modal features. Due to the heterogeneity of multimodal networks, optimizing a single objective will inevitably make the models prone to rely on specific modality while leaving other modalities under-optimized. On the other hand, simply expecting each modality to play a significant role in identifying all the rumors is also not appropriate as the multimodal fake news often involves tampering in only one modality. Therefore, how to find the genuine tampering on the per-sample basis becomes the key point to unlock the full power of each modality in a good collaborative manner. To address these issues, we propose a Uni-modal Event-agnostic Knowledge Distillation framework framework (UEKD), which aims to transfer knowledge contained in the fine-grained prediction from uni-modal teachers to the multimodal student model through modality-specific distillation. Specifically, we find that the uni-modal teachers simply trained on the whole training set are easy to memorize the event-specific noise information to make a correct but biased prediction, failing to reflect the genuine degree of tampering in each modality. To tackle this problem, we propose to train and validate the teacher models on different domains in training dataset through a cross-validation manner, as the predictions from the out-of-domain teachers can be regarded as event-agnostic knowledge without spurious connections with event-specific information. Finally, to balance the convergence speeds across modalities, we dynamically monitor the involvement of each modality during training, through which we could identify the more under-optimized modalities and re-weight the distillation loss accordingly. Our method could be served as a plug-and-play module for existing multimodal fake news detection backbones. Extensive experiments on three public datasets and four state-of-the-art fake news detection backbones show that our proposed method can improve the performance by a large margin.

Existing multimodal fake news detection methods based on traditional small models are prone to learn superficial features while struggling to perform deep reasoning and truly perceive fine-grained image-text consistency. Recently, fueled by Large Language Models and large-scale multimodal pretraining data, Large Vision-Language Models (LVLMs) have seen significant progress in these aspects. In our pilot study, domain-specific fine-tuned LVLM (FT-LVLM) Qwen2-VL-2B has achieved superior performance for multimodal fake news detection. However, we still find two weaknesses within it: (1) The model tends to rely only on textual modality while learning little from visual modality. (2) Directly outputting predictions hinders the model from exploiting the powerful reasoning capability of the LVLM. To address the above issues, we further propose a novel framework of Mixture-of-Experts Multi-View Large Visual-Language Models (MMVM) to explicitly disentangle multi-view features within the LVLM and adaptively fuse them for fake news detection. Specifically, we first extract representations of three distinct views, i.e., main branch, visual branch and reasoning branch, from the LVLM respectively. Then, considering that a given piece of fake news may not contain clues of equal importance in all the views, View-Aware Mixture-of-Experts networks (VAMoE) are proposed to dynamically reweight single-view representations and route the most informative features for the final prediction. Extensive experiments on four public datasets demonstrate the significant effectiveness and generalization ability of our proposed framework.