A Serial Two-Stage Framework for Robust Multimodal Fake News Detection via Adaptive Reasoning

KGEMD: knowledge-guided enhanced multimodal detection for fake news

Enhancing Fake News Detection using a BERT–BiLSTM Hybrid Framework

The rapid proliferation of digitally distributed news has made large-scale automated verification an urgent research priority, as deceptive content spreads across social platforms faster than manual reviewers can evaluate it. This work introduces a two-stage deep learning pipeline in which a pretrained BERT encoder generates rich contextual token embeddings that are subsequently processed by a Bidirectional Long Short-Term Memory (BiLSTM) network, enabling the classifier to exploit both sentence-level semantics and document-level narrative flow simultaneously. Experiments on the WELFake benchmark [1] following a standardised cleaning pipeline—lowercasing, URL stripping, tokenisation, stop-word removal, and WordNet lemmatisation— yielded 98.7 % accuracy and an F1-score of 0.986. This surpasses a BERT-only baseline by 3.1 percentage points and exceeds several previously published state-of-the-art results. The gains confirm that coupling transformer-based semantic representations with recurrent sequential modelling produces a measurable and reproducible improvement in misinformation detection. Planned extensions include multilingual evaluation and knowledge-distilled encoder replacements targeting real-time throughput constraints.

TERN: Type-Aware Evidence Reasoning for Multimodal Fake News Detection

Multimodal fake news detection remains challenging because deceptive posts exhibit heterogeneous manipulation patterns, while most existing methods still rely on a unified fusion strategy. This mismatch limits their ability to adapt to different evidence preferences across samples, encourages entanglement between deception cues and topical semantics, and weakens decision making when textual, visual, and cross-modal signals conflict. To address these issues, we propose TERN, a type-aware evidence reasoning network for multimodal fake news detection. TERN induces latent deception types from image-side multimodal features through prototype-based clustering, uses the induced assignments as structural priors for downstream veracity prediction, disentangles type-discriminative factors from semantic content, and performs type-conditioned hierarchical reasoning over text semantics, image authenticity, and cross-modal consistency. Experiments on MR2-Chinese, MR2-English, Weibo, and PHEME show that TERN achieves an average accuracy of 93.21% and an average F1 score of 91.09% while also improving Matthews correlation coefficient over representative multimodal baselines.

Fake News Identification Using Hybrid Transformer Ensemble Approach

A hybrid transformer-based ensemble model for automated fake news identification using the FakeNewsNet dataset is proposed and Experimental results show that the ensemble model achieves an accuracy of approximately 93%, outperforming the individual constituent models.

A Multi-Model Learning Framework for Fake News Detection on Social Media

The rapid proliferation of user-generated content on social media platforms has transformed the way information is created, shared, and consumed. While this democratization of information has clear benefits, it has also enabled the large-scale dissemination of fake news, which poses serious threats to public health, political stability, financial markets, and social cohesion. Automatic fake news detection has therefore become an active and urgent research problem in computer science. This paper proposes a multi-model learning framework that combines the complementary strengths of classical machine learning classifiers, deep sequential neural networks, and transformer-based contextual language models to detect fake news on social media. The framework integrates lexical, semantic, and contextual features through a weighted soft-voting ensemble that fuses predictions from a Support Vector Machine, a Bidirectional Long Short-Term Memory network, and a fine-tuned BERT encoder. Experiments were conducted on three widely used public benchmark datasets, namely LIAR, FakeNewsNet, and ISOT, comprising more than one hundred thousand labeled news statements and articles. The proposed ensemble achieved an accuracy of 96.4% and an F1-score of 96.1% on the ISOT dataset, outperforming each individual constituent model and several competitive baselines reported in the literature. Ablation studies confirm that the transformer component contributes the largest performance gain, while the ensemble fusion improves robustness and reduces variance across datasets. The results demonstrate that combining heterogeneous learning paradigms yields a more accurate and generalizable fake news detector than any single model in isolation. The paper also discusses computational cost, interpretability, and limitations, and outlines directions for multimodal and cross-lingual extensions.

MAR: Metacognitive Agentic Reasoning for Multimodal Fake News Detection

Multimodal fake news combining text and images has become increasingly prevalent, fueled by the rapid dissemination on social media. Existing approaches predominantly rely on supervised learning–driven small multimodal language models, yet they are constrained by the knowledge scope and logical reasoning capabilities limited by training data: their performance degrades significantly when encountering novel content not covered in the training data, and they typically provide only uninterpretable binary true/false predictions. To address these challenges, we propose the Metacognitive Agentic Reasoning for Multimodal Fake News Detection(MAR), a framework that integrates multi-agent metacognitive debate with external knowledge retrieval to enhance the accuracy, generalization, and interpretability of multimodal fake news detection. At its core, MAR introduces a news-domain informed and metacognition-inspired multi-agent reasoning mechanism: it first generates several prior pseudo-labels of the news domain, and defines the characteristics of two opposing agents (i.e., Believer and Skeptic) leveraging these pseudo-labels; then, the Believer and Skeptic go through a three-stage ''Draft – Self-Critique – Refine'' interactive debate simulating humans' learning behavior, called the metacognitive debate. Specifically, MAR first generates an initial evidence-grounded judgment (Draft), then critically reflects on their own and others' arguments (Self-Critique), and finally iteratively refines their conclusions by incorporating internal or external evidence (Refine). To improve factual reliability, the framework dynamically retrieves trustworthy external knowledge via web and reverse image searches, thereby mitigating the hallucinations inherent in large language models. Experiments show that MAR achieves state-of-the-art performance on two benchmarks and significantly outperforms existing methods in terms of generalization and interpretability. The source code is available at https://github.com/Averdgr/MAR_KDD.