Skip to content
Open access

Adaptive Semantic Feature Refinement for Explainable Fake News Detection Using Pretrained Transformers

Sudha Patel Shilpa Serasiya Sachi Bhavsar Maitri Rashmikant Sakarvadiya
Aug 2026 · International journal of computer information systems and industrial management applications · 0 citations

TL;DR

Results indicate that the proposed architecture successfully combines the process of refining the semantic features of text data while providing an explainable artificial intelligence solution for real-world applications of fake news detection.

Abstract

The online media has proliferated and become more accessible, so too has the ease with which misinformation can spread and be consumed and automated mechanisms to detect this form of online deception will be very important research targets moving forward. The implementation of new techniques from deep learning and transformer models pre-trained on large amounts of data has greatly improved the ability to detect misinformation, however many detectors are hindered by limitations on their ability to utilize semantic features and interpret the resulting predictions. In this study, we present our Adaptive Semantic Feature Refinement for Explainable Fake News Detection Utilizing Pre-Trained Transformers model, which uses a novel hybrid deep learning architecture that combines DeBERTa-v3, Bidirectional Gated Recurrent Unit (BiGRU), Multi-Head Self-Attention, and an Adaptive Semantic Refinement Module (ASRM) to produce higher-quality representations of the text used to classify fake news articles into binary categories. In addition, by applying SHAP (SHapley Additive exPlanations) values and Integrated Gradients to improve prediction transparency, we were able to produce both global and local explanations of the model's predictions. Our model was tested using a large corpus of fake news articles that included 682,661 articles, the resulting accuracy was 85.85%, with a ROC AUC statistic of .9294. Overall, results indicate that our proposed architecture successfully combines the process of refining the semantic features of text data while providing an explainable artificial intelligence solution for real-world applications of fake news detection.

Read PDF

Similar papers

Open access Jul 2026

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.

M. E. Babu, G. Sukanya · 0 citations
Conference Jul 2026

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.

Ardra P Namboodiri, Archa P S, Honey Mol O · 0 citations
Open access Aug 2026

4HAN: An Enhanced Neural Network for Fake News Detection using Hypergraph

Due to the fast proliferation of online news media and social networks, there is a significant increase in the dissemination of misleading data and fake news on digital platforms. Fake news detection is difficult due to the incapacity of deep learning models or conventional machine learning in comprehending semantic and higher-level interactions between news text, news author, publisher, and additional metadata. This results in decreased detection efficiency and enables the dissemination of misinformation that impacts societal behavior and public opinion. To address this problem, a Four-Level Hierarchical Attention Network (4HAN) that incorporates word-, sentence-, and headline-level attention, along with Hypergraph Convolution and Hypergraph Attention, is proposed using the LIAR dataset. By combining semantic feature extraction and relational dependency modeling, the 4HAN framework achieves better results. Results showed a detection accuracy rate of 96.00%, which beats multiple existing methodologies in fake news detection. These findings demonstrate that combining hierarchical attention with hypergraph learning provides more effective semantic and relational representation, leading to improved fake news detection performance and greater robustness in misinformation analysis.

Alpana A. Borse, Gajanan K. Kharate, N. Wasatkar · 0 citations
Open access Jul 2026

H-FuseNet: A Hybrid Multi-Representation Fusion Framework for Robust Misinformation Detection

H-FuseNet is proposed, a hybrid multi-representation fusion framework that combines pretrained transformer representations with deception-oriented handcrafted linguistic, stylistic, and semantic features that can improve the reliability of automated fake news detection systems.

A. Mahmood, Marek Sikora, M. Ather et al. · 0 citations
Open access 2026

Explainable Hybrid Transformer Model Based Fake News Detection System

This paper presents an extensive study of fake news detection, which involves manually curated linguistic features, classic machine learning techniques, DL, transfomer, and fusion multimodal, and examines the key multimodal benchmark, the Fakeddit dataset.

Anurag, Amandeep · 0 citations
Open access Aug 2026

Self-Attention Based Progressive Generative Adversarial Network optimized with Aquila Optimization Algorithm for Sentimental Analysis of Fake News Detection

SAPGAN-AOA-FND, a novel framework combining a self-attention-based progressive generative adversarial network (SAPGAN) enhanced using the Aquila optimization algorithm (AOA) tailored specifically for fake news identification, is proposed.

Shashikant Mahadu Bankar, Deepak Kumar Gupta · 0 citations