SemEval 2024 Task 4: Multilingual Detection of Persuasion Techniques in Memes
First Call for Participation
Memes that are part of a disinformation campaign achieve their goal of influencing social media users through a number of rhetorical and psychological techniques, such as causal oversimplification, name calling, appeal to fear, straw man, loaded language, and smears.
The goal of the shared task is to build models for identifying such techniques in the textual content of a meme as well as in a multimodal setting (in many cases requiring to make complex inferences using both textual and visual content). Additionally, the task offers a hierarchy of these techniques, which allows the use of more complex approaches when building models. Finally, there will be three surprise test datasets in different languages (a fourth one in English will be released as well), which will be revealed only at the final stages of the shared task.
Specifically, we offer the following subtasks:
Subtask 1 (multilabel classification problem; text only; multilingual test set): Given only the "textual content" of a meme, identify which of the 20 persuasion techniques, organized in a hierarchy, it uses.
Subtask 2a (multilabel classification problem; multimodal; multilingual test set): Given a meme, identify which of the 22 persuasion techniques, organized in a hierarchy, are used both in the textual and in the visual content of the meme
Subtask 2b (binary classification problem; multimodal): Given a meme (both the textual and the visual content), identify whether it contains a persuasion technique, or no technique.
The data is annotated with the following persuasion techniques:
Loaded Language; Name Calling/Labeling; Exaggeration/Minimization; Appeal to fear/prejudice; Flag-Waving; Slogans; Repetition; Doubt; Reductio ad Hitlerum; Obfuscation/Intentional Vagueness/Confusion; Smears; Glittering Generalities; Causal Oversimplification; Black-and-White Fallacy; Appeal to Authority; Bandwagon; Red Herring; Whataboutism; Thought-terminating Cliches; Straw Men.
We believe the tasks would be appealing to various NLP communities, including researchers working on sentiment analysis, fact-checking, argumentation mining, tagging, sequence modeling, as well as researchers working on image analysis in a multimodal scenario.
A live leaderboard will allow participants to track their progress on both tasks. All participants will be invited to submit a paper to the SemEval-2024 workshop.
Shared task website: https://propaganda.math.unipd.it/semeval2024task4
Competition dates: 4 September 2023 - 31 January 2024
Schedule October, Release of train labels
January 13, 2023 (tentative) Release of the gold labels of the dev set
January 20, 2024 (tentative) Release of the test set
January 31, 2024 at 23:59 (Anywhere on Earth) Test submission site closes
February 29, 2024 Paper Submission Deadline
April 1, 2024 Notification to authors
April 22, 2024 Camera ready papers due
June 16–21, 2024 SemEval 2024 workshop (co-located with NAACL 2024 in Mexico City, Mexico)
Task Organisers:
Dimitar Dimitrov, Sofia University "St. Kliment Ohrdiski"
Giovanni Da San Martino, University of Padova, Italy
Preslav Nakov, Mohamed bin Zayed University of Artificial Intelligence, UAE
Firoj Alam, Qatar Computing Research Institute, HBKU, Qatar
Maram Hasanain, Qatar Computing Research Institute, HBKU, Qatar
Abul Hasnat, Blackbird.ai
Fabrizio Silvestri, Sapienza University, Rome