Health Misinformation Detection in Web Content

Abstract

In recent years, we have witnessed the proliferation of large amounts of online content generated directly by users with virtually no form of external control, leading to the possible spread of misinformation. The search for effective solutions to this problem is still ongoing, and covers different areas of application, from opinion spam to fake news detection. A more recently investigated scenario, despite the serious risks that incurring disinformation could entail, is that of the online dissemination of health information. Early approaches in this area focused primarily on userbased studies applied to Web page content. More recently, automated approaches have been developed for both Web pages and social media content, particularly with the advent of the COVID-19 pandemic. These approaches are primarily based on handcrafted features extracted from online content in association with Machine Learning. In this scenario, we focus on Web page content, where there is still room for research to study structural-, content-and context-based features to assess the credibility of Web pages. Therefore, this work aims to study the effectiveness of such features in association with a deep learning model, starting from an embedded representation of Web pages that has been recently proposed in the context of phishing Web page detection, i.e., Web2Vec.

The Explainability of Transformers: Current Status and Directions

“…In the following, we report the task for which explainability was sought through SHAP. Upadhyay et al proposed a new model for fake health news detection [131]. Abbruzzese et al proposed a new architecture for OCR anomaly detection and correction [132].…”

Online health search via multi-dimensional information quality assessment based on deep language models

“…Recently, health misinformation-related topics, like "vaccine" or "the relationship between coronavirus and 5G", on social media platforms, such as Twitter and Instagram [13][14][15], and across countries [16] have gained scientific interest. Thus, new credibility-centred search methods and assessment measures are crucial for tackling open issues in health-related information retrieval [17]. In the past few years, information retrieval shared tasks, such as the Text REtrieval Conference (TREC) and the Conference and Labs of the Evaluation Forum (CLEF), have started evaluating quality-based systems for health corpora [18,19].…”

“…In the absence of sufficiently large training data, classical bag-of-words-based methods tend to outperform neural network approaches in the detection of health-related misinformation [26]. On the other hand, a combination of classical handcrafted feature-based methods and deep learning-based methods can be effective when customized for health misinformation detection in Web content [17]. Multi-stage ranking systems [27,28] coupled with a label prediction model [29] improve upon bag-of-word-based baselines and have set the state-of-the-art [28] for separating helpful from harmful Web content [30].…”

“…While classical bag-of-words-based methods outperform neural network approaches in detecting health-related misinformation when training data is limited [29], more advanced approaches are needed for Web content. Specifically, research has proven the effectiveness of a hybrid approach that integrates classical handcrafted features with deep learning [18]. Further to this, multi-stage ranking systems [30,31] have been proposed, which couple the system with a label prediction model or employ T5 [33] to re-rank BM25 results.…”

“…Topics related to misinformation, such as 'vaccine' or 'the relationship between coronavirus and 5G' have gained scientific interest across social media platforms like Twitter and Instagram [14][15][16] and among various countries [17]. Thus, the development of new credibility-centered search methods and assessment measures is crucial to address the pressing challenges in health-related information retrieval [18].…”