See the MediaEval 2021 webpage for information on how to register and participate.
The FakeNews Detection Task offers three fake news detection subtasks on COVID-19-related conspiracy theories. The first subtask includes text-based fake news detection, the second subtask targets the detection of conspiracy theory topics, and the third subtask combines topic and conspiracy detection. All subtasks are related to misinformation disseminated in the context of the long-lasting COVID-19 crisis. We focus on conspiracy theories that assume some kind of nefarious actions by governments or other actors related to CODID-19, such as intentionally spreading the pandemic, lying about the nature of the pandemic, or using vaccines that have some hidden functionality and purpose.
Text-Based Misinformation Detection: In this subtask, the participants receive a dataset consisting of tweet text blocks in English related to COVID-19 and various conspiracy theories. The participants are encouraged to build a multi-class classifier that can flag whether a tweet promotes/supports or discusses at least one (or many) of the conspiracy theories. In the case if the particular tweet promotes/supports one conspiracy theory and just discusses another, the result of the detection for the particular tweet is experted to be equal to “stronger” class: promote/support in the given sample.
Text-Based Conspiracy Theories Recognition: In this subtask, the participants receive a dataset consisting of tweet text blocks in English related to COVID-19 and various conspiracy theories. The main goal of this subtask is to build a detector that can detect whether a text in any form mentions or refers to any of the predefined conspiracy topics.
Text-Based Combined Misinformation and Conspiracies Detection: In this subtask, the participants receive a dataset consisting of tweet text blocks in English related to COVID-19 and various conspiracy theories. The goal of this subtask is to build a complex multi-labelling multi-class detector that for each topic from a list of predefined conspiracy topics can predict whether a tweet promotes/supports or just discusses that particular topic.
Digital wildfires, i.e., fast-spreading inaccurate, counterfactual, or intentionally misleading information, can quickly permeate public consciousness and have severe real-world implications, and they are among the top global risks in the 21st century. While a sheer endless amount of misinformation exists on the internet, only a small fraction of it spreads far and affects people to a degree where they commit harmful and/or criminal acts in the real world. The COVID-19 pandemic has severely affected people worldwide, and consequently, it has dominated world news for months. Thus, it is no surprise that it has also been the topic of a massive amount of misinformation, which was most likely amplified by the fact that many details about the virus were unknown at the start of the pandemic. This task aims at the development of methods capable of detecting such misinformation. Since many different misinformation narratives exist, such methods must be capable of distinguishing between them. For that reason we consider a variety of well-known conspiracy theories related to COVID-19.
The task is of interest to researchers in the areas of online news, social media, multimedia analysis, multimedia information retrieval, natural language processing, and meaning understanding and situational awareness to participate in the challenge.
The dataset contains several sets of tweet texts mentioning Corona Virus and different conspiracy theories. The dataset set consists of only English language posts and it contains a variety of long tweets with neutral, positive, negative, and sarcastic phrasing. The datasets is not balanced with respect to the number of samples of conspiracy-promoting and other tweets, and the number of tweets per each conspiracy class. The dataset items have been collected from Twitter during a period between 20th of January 2020 and 31st of July 2021, by searching for the Corona-virus-related keywords (e.g., “corona”, “COVID-19”, etc.) inside the tweets’ text, followed by a search for keywords related to the conspiracy theories. Since not all tweets are available online, the partipants will be provided a full-text set of already downloaded tweets. In order to be compliant with the Twitter Developer Policy, only the members of the participants’ participating temas are allowed to access and use the provided dataset. Distribution, publication, sharing and any form of usage of the provided data apart of the research purposes within the FakeNews task is strictly prohibited. A copy of the dataset in form of Tweet ID and annotations will be published after the end of MediaEval 2021.
The ground truth for the provided dataset was created by the team of well-motivated students and researchers using overlapping annotation process with the following cross-validation and verification by an independent assisting team.
Evaluation will be performed using standard implementation of the multi-class generalization of the Matthews correlation coefficient (MCC, https://scikit-learn.org/stable/modules/generated/sklearn.metrics.matthews_corrcoef.html) computed on the optimally threshold conspiracy promoting probabilities (threshold that yields the best MCC score).
 Nyhan, Brendan, and Jason Reifler. 2015. Displacing misinformation about events: An experimental test of causal corrections. Journal of Experimental Political Science 2, no. 1, 81-93.
Twitter data collection and analysis
 Burchard, Luk, Daniel Thilo Schroeder, Konstantin Pogorelov, Soeren Becker, Emily Dietrich, Petra Filkukova, and Johannes Langguth. 2020. A Scalable System for Bundling Online Social Network Mining Research. In 2020 Seventh International Conference on Social Networks Analysis, Management and Security (SNAMS), IEEE, 1-6.
 Schroeder, Daniel Thilo, Konstantin Pogorelov, and Johannes Langguth. 2019. FACT: a Framework for Analysis and Capture of Twitter Graphs. In 2019 Sixth International Conference on Social Networks Analysis, Management and Security (SNAMS), IEEE, 134-141.
 Achrekar, Harshavardhan, Avinash Gandhe, Ross Lazarus, Ssu-Hsin Yu, and Benyuan Liu. 2011. Predicting flu trends using twitter data. In 2011 IEEE conference on computer communications workshops (INFOCOM WKSHPS), IEEE, 702-707.
 Chen, Emily, Kristina Lerman, and Emilio Ferrara. 2020. Covid-19: The first public coronavirus twitter dataset. arXiv preprint arXiv:2003.07372.
 Kouzy, Ramez, Joseph Abi Jaoude, Afif Kraitem, Molly B. El Alam, Basil Karam, Elio Adib, Jabra Zarka, Cindy Traboulsi, Elie W. Akl, and Khalil Baddour. 2020. Coronavirus goes viral: quantifying the COVID-19 misinformation epidemic on Twitter. Cureus 12, no. 3.
Natural language processing
 Bourgonje, Peter, Julian Moreno Schneider, and Georg Rehm. 2017. From clickbait to fake news detection: an approach based on detecting the stance of headlines to articles. In Proceedings of the 2017 EMNLP Workshop: Natural Language Processing meets Journalism, 84-89.
 Imran, Muhammad, Prasenjit Mitra, and Carlos Castillo. 2016. Twitter as a lifeline: Human-annotated twitter corpora for NLP of crisis-related messages. arXiv preprint arXiv:1605.05894.
 Liu, Chuang, Xiu-Xiu Zhan, Zi-Ke Zhang, Gui-Quan Sun, and Pak Ming Hui. 2015. How events determine spreading patterns: information transmission via internal and external influences on social networks. New Journal of Physics 17, no. 11.
Online news sources analysis
 Pogorelov, Konstantin, Daniel Thilo Schroeder, Petra Filkukova, and Johannes Langguth. 2020. A System for High Performance Mining on GDELT Data. In 2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), IEEE, 1101-1111.
This work was funded by the Norwegian Research Council under contracts #272019 and #303404 and has benefited from the Experimental Infrastructure for Exploration of Exascale Computing (eX3), which is financially supported by the Research Council of Norway under contract #270053. We also acknowledge support from Michael Kreil in the collection of Twitter data.