Dear Colleagues,
We are excited to announce that the Leaderboard for SemEval 2025 Task 10:
Multilingual Characterization and Extraction of Narratives from Online News
is officially open!
Task Overview
The task focuses on analyzing multilingual news articles in two critical
domains—Ukraine-Russia War and Climate Change—and is subdivided into three
subtasks:
1.
Entity Framing
-
Goal: Assign fine-grained roles (e.g., protagonist, antagonist, or
innocent) to mentions of named entities in a news article.
-
Task Type: Multi-label, multi-class text-span classification.
2.
Narrative Classification
-
Goal: Assign appropriate sub-narrative labels to a news article using
a two-level taxonomy of narratives.
-
Task Type: Multi-label, multi-class document classification.
3.
Narrative Extraction
-
Goal: Generate a free-text explanation (up to 80 words) for the
dominant narrative in an article, grounded in evidence from the text.
-
Task Type: Text-to-text generation.
This task provides an opportunity to push the boundaries of multilingual NLP
and tackle challenges related to narrative understanding and extraction.
Important Dates
-
Development Set Available: November 11, 2024
-
Leaderboard Open: November 11, 2024
-
Test Set Release: January, 2025
-
Submission Deadline: January, 2025 (keep monitoring the website for
precise dates)
How to Participate
1.
Visit the official task website:
https://propaganda.math.unipd.it/semeval2025task10/
2.
Register your team to get access to the data and begin evaluating your
models.
3.
Submit your predictions to the leaderboard to compare with other
participants.
This is a great opportunity to contribute to cutting-edge research in
multilingual
narrative analysis and to engage with the SemEval community.
We look forward to your participation!
Best regards,
The SemEval 2025 Task 10 Organizers
_______________________________________________
Corpora mailing list -- [email protected]
https://list.elra.info/mailman3/postorius/lists/corpora.list.elra.info/
To unsubscribe send an email to [email protected]
