Semi-supervised learning and graph neural networks for fake news detection

Filtering, vetting, and verifying digital information is an area of core interest in information science. Online fake news is a specific type of digital misinformation that poses serious threats to democratic institutions, misguides the public, and can lead to radicalization and violence. Hence, fake news detection is an important problem for information science research. While there have been multiple attempts to identify fake news, most of such efforts have focused on a single modality (e.g., only text‐based or only visual features). However, news articles are increasingly framed as multimodal news stories, and hence, in this work, we propose a multimodal approach combining text and visual analysis of online news stories to automatically detect fake news. Drawing on key theories of information processing and presentation, we identify multiple text and visual features that are associated with fake or credible news articles. We then perform a predictive analysis to detect features most strongly associated with fake news. Next, we combine these features in predictive models using multiple machine‐learning techniques. The experimental results indicate that a multimodal approach outperforms single‐modality approaches, allowing for better fake news detection.

Recently, fake news has been widely spread through the Internet due to the increased use of social media for communication. Fake news has become a significant concern due to its harmful impact on individual attitudes and the community’s behavior. Researchers and social media service providers have commonly utilized artificial intelligence techniques in the recent few years to rein in fake news propagation. However, fake news detection is challenging due to the use of political language and the high linguistic similarities between real and fake news. In addition, most news sentences are short, therefore finding valuable representative features that machine learning classifiers can use to distinguish between fake and authentic news is difficult because both false and legitimate news have comparable language traits. Existing fake news solutions suffer from low detection performance due to improper representation and model design. This study aims at improving the detection accuracy by proposing a deep ensemble fake news detection model using the sequential deep learning technique. The proposed model was constructed in three phases. In the first phase, features were extracted from news contents, preprocessed using natural language processing techniques, enriched using n-gram, and represented using the term frequency–inverse term frequency technique. In the second phase, an ensemble model based on deep learning was constructed as follows. Multiple binary classifiers were trained using sequential deep learning networks to extract the representative hidden features that could accurately classify news types. In the third phase, a multi-class classifier was constructed based on multilayer perceptron (MLP) and trained using the features extracted from the aggregated outputs of the deep learning-based binary classifiers for final classification. The two popular and well-known datasets (LIAR and ISOT) were used with different classifiers to benchmark the proposed model. Compared with the state-of-the-art models, which use deep contextualized representation with convolutional neural network (CNN), the proposed model shows significant improvements (2.41%) in the overall performance in terms of the F1score for the LIAR dataset, which is more challenging than other datasets. Meanwhile, the proposed model achieves 100% accuracy with ISOT. The study demonstrates that traditional features extracted from news content with proper model design outperform the existing models that were constructed based on text embedding techniques.

Nowadays, the dissemination of news information has become more rapid, liberal, and open to the public. People can find what they want to know more and more easily from a variety of sources, including traditional news outlets and new social media platforms. However, at a time when our lives are glutted with all kinds of news, we cannot help but doubt the veracity and legitimacy of these news sources; meanwhile, we also need to guard against the possible impact of various forms of fake news. To combat the spread of misinformation, more and more researchers have turned to natural language processing (NLP) approaches for effective fake news detection. However, in the face of increasingly serious fake news events, existing detection methods still need to be continuously improved. This study proposes a modified proof-of-concept model named NER-SA, which integrates natural language processing (NLP) and named entity recognition (NER) to conduct the in-domain and cross-domain analysis of fake news detection with the existing three datasets simultaneously. The named entities associated with any particular news event exist in a finite and available evidence pool. Therefore, entities must be mentioned and recognized in this entity bank in any authentic news articles. A piece of fake news inevitably includes only some entitlements in the entity bank. The false information is deliberately fabricated with fictitious, imaginary, and even unreasonable sentences and content. As a result, there must be differences in statements, writing logic, and style between legitimate news and fake news, meaning that it is possible to successfully detect fake news. We developed a mathematical model and used the simulated annealing algorithm to find the optimal legitimate area. Comparing the detection performance of the NER-SA model with current state-of-the-art models proposed in other studies, we found that the NER-SA model indeed has superior performance in detecting fake news. For in-domain analysis, the accuracy increased by an average of 8.94% on the LIAR dataset and 19.36% on the fake or real news dataset, while the F1-score increased by an average of 24.04% on the LIAR dataset and 19.36% on the fake or real news dataset. In cross-domain analysis, the accuracy and F1-score for the NER-SA model increased by an average of 28.51% and 24.54%, respectively, across six domains in the FakeNews AMT dataset. The findings and implications of this study are further discussed with regard to their significance for improving accuracy, understanding context, and addressing adversarial attacks. The development of stylometric detection based on NLP approaches using NER techniques can improve the effectiveness and applicability of fake news detection.

Due to the widespread use of social media, people are exposed to fake news and misinformation. Spreading fake news has adverse effects on both the general public and governments. This issue motivated researchers to utilize advanced natural language processing concepts to detect such misinformation in social media. Despite the recent research studies that only focused on semantic features extracted by deep contextualized text representation models, we aim to show that content-based feature engineering can enhance the semantic models in a complex task like fake news detection. These features can provide valuable information from different aspects of input texts and assist our neural classifier in detecting fake and real news more accurately than using semantic features. To substantiate the effectiveness of feature engineering besides semantic features, we proposed a deep neural architecture in which three parallel convolutional neural network (CNN) layers extract semantic features from contextual representation vectors. Then, semantic and content-based features are fed to a fully connected layer. We evaluated our model on an English dataset about the COVID-19 pandemic and a domain-independent Persian fake news dataset (TAJ). Our experiments on the English COVID-19 dataset show 4.16% and 4.02% improvement in accuracy and f1-score, respectively, compared to the baseline model, which does not benefit from the content-based features. We also achieved 2.01% and 0.69% improvement in accuracy and f1-score, respectively, compared to the state-of-the-art results reported by Shifath et al. (A transformer based approach for fighting covid-19 fake news, arXiv preprint arXiv:2101.12027, 2021). Our model outperformed the baseline on the TAJ dataset by improving accuracy and f1-score metrics by 1.89% and 1.74%, respectively. The model also shows 2.13% and 1.6% improvement in accuracy and f1-score, respectively, compared to the state-of-the-art model proposed by Samadi et al. (ACM Trans Asian Low-Resour Lang Inf Process, https://doi.org/10.1145/3472620, 2021).

Due to the widespread use of social networks by people of all ages, distinguishing between fake and real news on these platforms has become a significant challenge in today's world. Individuals who disseminate fake news on social networks often aim to achieve various commercial, political, and economic goals. Therefore, identifying and distinguishing real news from fake news is crucial in addressing this issue. The objective of this research is to present an intelligent model for detecting fake news using a news propagation tree in the social network Twitter. The dataset used in this study is sourced from the political news section of the Fake News Net website. Initially, a news propagation tree was constructed for both real and fake news using this dataset, followed by the development of features from structural, temporal, and syntactic perspectives based on the news propagation tree. Finally, machine learning algorithms were employed to build a model for predicting fake and real news. The results indicated that among the algorithms used for modeling, the Random Forest algorithm, with an accuracy of 75.8%, was the best model for distinguishing fake news from real news.

Fake news detection within online social media using supervised artificial intelligence algorithms

Fake news have attracted attention of general public because of the influence they can exert on important activities of society, such as elections. Efforts have been made to detect them, but usually they rely on human labour fact-checking, what can be costly and time consuming. Computational approaches have typically relied on supervised learning models, in which a model is trained based on fake and true news samples. Such approach allows a large amount of news to be classified in a short time, but it demands datasets labelled with positive and negative instances. Our work proposes to detect fake news by training a model with only fake samples in the training dataset, through One-Class Classification (OCC). We compare a novel algorithm, called DCDistanceOCC, to others published in literature, and got similar, or even better, results. The case study is the Brazilian politics scenario starting at the 2018 general elections on Twitter and WhatsApp. These two platforms were a fertile ground to fake news proliferation. We also evaluated the models over another available dataset from literature. To the best of our knowledge, this is the first paper to identify fake news using an OCC approach and also the first one to provide Portuguese-based WhatsApp and Twitter datasets with fake news.

In perceptual decision-making, it is well-known that low confidence in a decision promotes information seeking, which increases subsequent decision accuracy. It is unclear, however, whether these mechanisms generalize to more complex domains such as fake news detection. Here, we tested whether confidence also drives information seeking in the context of fake news, whether seeking information improves fake news detection and sharing intentions, and whether we find confidence-driven confirmation biases. In an online study, 314 US citizens indicated whether news headlines were true or fake and could request additional information before finalizing their decision. First, in line with findings from perceptual decision-making, confidence was the main driver of information seeking for both fake and true news. Yet critically, whereas confidence lawfully scaled with decision accuracy for true news items, we found a U-shaped relationship for fake news items: decision accuracy was lowest for intermediate confidence, indicating that for fake headlines confidence is not a reliable signal of headline veracity. Second, while information seeking increased decision accuracy for true news items, for fake news items decision accuracy remained at chance level even after information seeking. Sharing intentions for fake news remained equally unaffected by information seeking, and mostly increased for belief-congruent true items. Finally, we found a confidence-driven confirmation bias for beliefs, but not for sharing intentions. Together, our results clarify that while confidence-driven information seeking adaptively increases decision accuracy for true news items, it is entirely ineffective in the context of fake news.

Fake news gains has gained significant momentum, strongly motivating the need for fake news research. Many fake news detection approaches have thus been proposed, where most of them heavily rely on news content. However, networkbased clues revealed when analyzing news propagation on social networks is an information that has hardly been comprehensively explored or used for fake news detection. We bridge this gap by proposing a network-based pattern-driven fake news detection approach. We aim to study the patterns of fake news in social networks, which refer to the news being spread, spreaders of the news and relationships among the spreaders. Empirical evidence and interpretations on the existence of such patterns are provided based on social psychological theories. These patterns are then represented at various network levels (i.e., node-level, ego-level, triad-level, community-level and the overall network) for being further utilized to detect fake news. The proposed approach enhances the explainability in fake news feature engineering. Experiments conducted on real-world data demonstrate that the proposed approach can outperform the state of the arts.

Fake news detection on job advertisements has grabbed the attention of many researchers over past decade. Various classifiers such as Support Vector Machine (SVM), XGBoost Classifier and Random Forest (RF) methods are greatly utilized for fake and real news detection pertaining to job advertisement posts in social media. Bi-Directional Long Short-Term Memory (Bi-LSTM) classifier is greatly utilized for learning word representations in lower-dimensional vector space and learning significant words word embedding or terms revealed through Word embedding algorithm. The fake news detection is greatly achieved along with real news on job post from online social media is achieved by Bi-LSTM classifier and thereby evaluating corresponding performance. The performance metrics such as Precision, Recall, F1-score, and Accuracy are assessed for effectiveness by fraudulency based on job posts. The outcome infers the effectiveness and prominence of features for detecting false news. .

Easy access, variety of content, and fast widespread interactions are some of the reasons making social media increasingly popular. However, this rise has also enabled the propagation of fake news, text published by news sources with an intent to spread misinformation and sway beliefs. Detecting it is an important and challenging problem to prevent large scale misinformation and maintain a healthy society. We view fake news detection as reasoning over the relations between sources, articles they publish, and engaging users on social media in a graph framework. After embedding this information, we formulate inference operators which augment the graph edges by revealing unobserved interactions between its elements, such as similarity between documents’ contents and users’ engagement patterns. Our experiments over two challenging fake news detection tasks show that using inference operators leads to a better understanding of the social media framework enabling fake news spread, resulting in improved performance.

Nowadays, with the advancement of technology, the use of news sources has also undergone a great evolution. News sources have constantly evolved from past to present, ranging from magazines to radios, from newspapers to televisions. The fact that it has become so easy to access news has caused society to pay more attention to fake news. Fake news has the ability to influence society through news sources such as social media, which can reach wider audiences with the development of technology. The difficulties of users in accessing accurate and reliable sources in this information flow that shapes their daily lives increases the potential for the spread of fake news, and it becomes increasingly difficult to distinguish between real and fake news. In this study, classification models for fake news detection were designed using machine learning algorithms. The dataset, which includes fake and real news examples, contains 42,000 examples. Each class, including fake and real samples, contains 22,000 sample data. In order to increase data quality, accuracy and usability, preprocessing methods were applied to the data set. The removal of numbers, stop words, and html tags was done in the pre-processing step to remove unnecessary information from the text. Models were created for fake news detection with singular and ensemble classification algorithms. Performance evaluation of the models was performed using 5-fold cross-validation. In the performance comparisons of the models, values such as accuracy, sensitivity, specificity, tp rate and fp rate were calculated. The highest performance results were observed in the random forest classification algorithm with an accuracy rate of 76%.

Classifier algorithms are a subfield of data mining and play an integral role in finding patterns and relationships within large datasets. In recent years, fake news detection has become a popular area of data mining for several important reasons, including its negative impact on decision-making and its virality within social networks. In the past, traditional fake news detection has relied primarily on information context, while modern approaches rely on auxiliary information to classify content. Modelling with machine learning and natural language processing can aid in distinguishing between fake and real news. In this research, we mine data from Reddit, the popular online discussion forum and social news aggregator, and measure machine learning classifiers in order to evaluate each algorithm’s accuracy in detecting fake news using only a minimal subset of data.

In the extensive growth of social media platforms, fake news is false information spread among people around the world. Fake news weakens public trust and societal stability; it has negative impacts that necessitate for fake news. In order to enhance the public trust and improve the social media users, fake news detection is crucial, which is a complex and multi-dimensional task due to the diverse characteristics of fake news. More researchers have explored several deep learning and machine learning approaches for fake news detection approaches yet the existing binary classification mechanism is not sufficient for handling imbalanced training data to enhance poor performance in the fake news detection process also it affects the interpretability and consumes high computational duration. With the aim of rectifying these complexities, an automated fake news detection mechanism is proposed in this research work to ultimately recognise unreliable information for improving social media users’ effectiveness. Initially, the necessary data are taken from the publicly available datasets, such as the WELFake dataset and Fake News dataset; it is fed into the preprocessing phase. Here, the punctuation, special characters, redundant content and inappropriate data are mitigated, which can generate cleaned data. Then, the preprocessed data is given to the feature extraction process to efficiently extract the relevant informative features for enhancing better detection outcomes. Here, the Bidirectional Encoder Representations from Transformers (BERT) is initially used to retrieve the first set of features, the Recurrent Neural Network (RNN) extracts the temporal features and the Convolution Neural Network (CNN) is used to extract the spatial features. Subsequently, these features are fused in the weighted feature selection process, where the weights are quickly tuned through the Improved Horse Herd Optimisation Algorithm (IHHOA). Further, it is passed to the Residual Convolution Bidirectional Long Short Term Memory (RConv-Bi-LSTM) technique for significantly detecting fake news in a limited duration. The entire detection process is estimated with several performance metrics and contrasted over various conventional methods to generate optimal outcomes. Here, the designed framework’s FM value (dataset 1) is enhanced over 22.9% DNN, 8.7% CNN, 17.6% LSTM and 7% Bi-LSTM methods (BERT features) to show its better performance over the existing approaches.

Fake news is becoming a massive issue from being used to manipulate elections to spread chaos and to being darn right annoying. Fake news has become a thriving business where many companies and state agencies offer services to spread fake news deliberately to benefit their clients. Hence new techniques are required to identify and detect fake news. In this paper, we propose a novel 3-phase approach for the identification of fake news. We extracted features like WHOIS and DNS from domain information and used for the model construction. We also combine psychological factors of both the people trying to spread the fake news and the general population. We train our model on journalist labelled data and achieved more than 80% accuracy. We are also able to increase the precision and recall of fake and genuine news detection.