The COVID-19 social media infodemic

Scientific Reports, Год журнала: 2020, Номер 10(1)

Опубликована: Окт. 6, 2020

We address the diffusion of information about COVID-19 with a massive data analysis on Twitter, Instagram, YouTube, Reddit and Gab. analyze engagement interest in topic provide differential assessment evolution discourse global scale for each platform their users. fit spreading epidemic models characterizing basic reproduction numbers $R_0$ social media platform. Moreover, we characterize from questionable sources, finding different volumes misinformation However, both reliable sources do not present patterns. Finally, platform-dependent numerical estimates rumors' amplification.

Язык: Английский

---

BERT Models for Arabic Text Classification: A Systematic Review

Applied Sciences, Год журнала: 2022, Номер 12(11), С. 5720 - 5720

Опубликована: Июнь 4, 2022

Bidirectional Encoder Representations from Transformers (BERT) has gained increasing attention researchers and practitioners as it proven to be an invaluable technique in natural languages processing. This is mainly due its unique features, including ability predict words conditioned on both the left right context, pretrained using plain text corpus that enormously available web. As BERT more interest, models were introduced support different languages, Arabic. The current state of knowledge practice applying Arabic classification limited. In attempt begin remedying this gap, review synthesizes have been applied classification. It investigates differences between them compares their performance. also examines how effective they are compared original English models. concludes by offering insight into aspects need further improvements future work.

Язык: Английский

---

COVID-19: Detecting Government Pandemic Measures and Public Concerns from Twitter Arabic Data Using Distributed Machine Learning

International Journal of Environmental Research and Public Health, Год журнала: 2021, Номер 18(1), С. 282 - 282

Опубликована: Янв. 1, 2021

Today’s societies are connected to a level that has never been seen before. The COVID-19 pandemic exposed the vulnerabilities of such an unprecedently world. As 19 November 2020, over 56 million people have infected with nearly 1.35 deaths, and numbers growing. state-of-the-art social media analytics for COVID-19-related studies understand various phenomena happening in our environment limited require many more studies. This paper proposes software tool comprising collection unsupervised Latent Dirichlet Allocation (LDA) machine learning other methods analysis Twitter data Arabic aim detect government measures public concerns during pandemic. is described detail, including its architecture, five components, algorithms. Using tool, we collect dataset 14 tweets from Kingdom Saudi Arabia (KSA) period 1 February 2020 June 2020. We 15 six macro-concerns (economic sustainability, etc.), formulate their information-structural, temporal, spatio-temporal relationships. For example, able timewise progression events discussions on cases mid-March first curfew 22 March, financial loan incentives increased quarantine March–April, reduced mobility levels 24 March onwards, blood donation shortfall late government’s 9 billion SAR (Saudi Riyal) salary 3 April, lifting ban daily prayers mosques 26 May, finally return normal 29 May These findings show effectiveness detecting important events, measures, concerns, information both time space no earlier knowledge about them.

Язык: Английский

---

The CLEF-2021 CheckThat! Lab on Detecting Check-Worthy Claims, Previously Fact-Checked Claims, and Fake News

Lecture notes in computer science, Год журнала: 2021, Номер unknown, С. 639 - 649

Опубликована: Янв. 1, 2021

Язык: Английский

---

Overview of CheckThat! 2020: Automatic Identification and Verification of Claims in Social Media

Lecture notes in computer science, Год журнала: 2020, Номер unknown, С. 215 - 236

Опубликована: Янв. 1, 2020

Язык: Английский

---

Ensemble Classifiers for Arabic Sentiment Analysis of Social Network (Twitter Data) towards COVID-19-Related Conspiracy Theories

Applied Computational Intelligence and Soft Computing, Год журнала: 2022, Номер 2022, С. 1 - 10

Опубликована: Янв. 13, 2022

Sentiment analysis has recently become increasingly important with a massive increase in online content. It is associated the of textual data generated by social media that can be easily accessed, obtained, and analyzed. With emergence COVID-19, most published studies related to COVID-19’s conspiracy theories were surveys on people's sentiments opinions studied impact pandemic their lives. Just few utilized sentiment using machine learning approach. These focused more Twitter tweets English language did not pay attention other languages such as Arabic. This study proposes model analyze Arabic from Twitter. In this model, we apply Word2Vec for word embedding which formed main source features. Two pretrained continuous bag-of-words (CBOW) models are investigated, Naïve Bayes was used baseline classifier. Several single-based ensemble-based classifiers have been without SMOTE (synthetic minority oversampling technique). The experimental results show applying an ensemble achieved good improvement average F1 score compared classifier (single-based ensemble-based) SMOTE.

Язык: Английский

---

Overview of the WANLP 2022 Shared Task on Propaganda Detection in Arabic

Опубликована: Янв. 1, 2022

Propaganda is defined as an expression of opinion or action by individuals groups deliberately designed to influence opinions actions other with reference predetermined ends and this achieved means well-defined rhetorical psychological devices. Currently, propaganda (or persuasion) techniques have been commonly used on social media manipulate mislead users. Automatic detection from textual, visual, multimodal content has studied recently, however, major such efforts are focused English language content. In paper, we propose a shared task detecting for Arabic textual We done pilot annotation 200 tweets, which plan extend 2,000 covering diverse topics. hope that the will help in building community detection. The dataset be made publicly available, can future studies.

Язык: Английский

---

The state of human-centered NLP technology for fact-checking

Information Processing & Management, Год журнала: 2022, Номер 60(2), С. 103219 - 103219

Опубликована: Дек. 21, 2022

Язык: Английский

---

Detecting Harmful Content on Online Platforms: What Platforms Need vs. Where Research Efforts Go

ACM Computing Surveys, Год журнала: 2023, Номер 56(3), С. 1 - 17

Опубликована: Июнь 7, 2023

The proliferation of harmful content on online platforms is a major societal problem, which comes in many different forms, including hate speech, offensive language, bullying and harassment, misinformation, spam, violence, graphic content, sexual abuse, self-harm, others. Online seek to moderate such limit harm, comply with legislation, create more inclusive environment for their users. Researchers have developed methods automatically detecting often focusing specific sub-problems or narrow communities, as what considered depends the platform context. We argue that there currently dichotomy between types curb, research efforts are detect content. thus survey existing well moderation policies by this light suggest directions future work.

Язык: Английский

---

Zero-Shot Rumor Detection with Propagation Structure via Prompt Learning

Proceedings of the AAAI Conference on Artificial Intelligence, Год журнала: 2023, Номер 37(4), С. 5213 - 5221

Опубликована: Июнь 26, 2023

The spread of rumors along with breaking events seriously hinders the truth in era social media. Previous studies reveal that due to lack annotated resources, presented minority languages are hard be detected. Furthermore, unforeseen not involved yesterday's news exacerbate scarcity data resources. In this work, we propose a novel zero-shot framework based on prompt learning detect falling different domains or languages. More specifically, firstly represent rumor circulated media as diverse propagation threads, then design hierarchical encoding mechanism learn language-agnostic contextual representations for both prompts and data. To further enhance domain adaptation, model domain-invariant structural features from incorporate position influential community response. addition, new virtual response augmentation method is used improve training. Extensive experiments conducted three real-world datasets demonstrate our proposed achieves much better performance than state-of-the-art methods exhibits superior capacity detecting at early stages.

Язык: Английский

---