AraGenre 2026 shared task - First Call for Participation

25 views
Skip to first unread message

El-Haj, Mo

unread,
May 18, 2026, 2:51:46 AMMay 18
to SIGARAB: Special Interest Group on Arabic Natural Language Processing
Dear all,
I'm pleased to announce our new shared task at ArabicNLP @ EMNLP 2026 in Budapest, Hungry, October 24-29, 2026.
AraGenre: A Hierarchical Definition-Guided Arabic Genre Classification Shared Task
AraGenre is a challenging shared task designed to test how NLP systems generalise to the unseen through:
• Few-shot training and development data
• Cross-lingual genre definitions guiding Arabic classification
• A large hidden test set of naturally occurring Arabic texts
• Unseen genre-definition combinations
The task addresses a wider challenge in NLP: building methods that can generalise from limited data and support languages and varieties that remain underrepresented in current NLP research. 
Registration now open:
https://forms.gle/RAi7SpYoRD5RgUnP6
Organisers:
- Mo El-Haj, VinUniversty, Vietnam / Lancaster University, UK.
- Saad Ezzini, KFUPM, KSA
- Mustafa Jarrar, HBKU, Qatar
- Shadi Abudalfa, KFUPM, KSA

El-Haj, Mo

unread,
Jun 17, 2026, 6:45:57 AM (6 days ago) Jun 17
to SIGARAB: Special Interest Group on Arabic Natural Language Processing
📢 AraGenre Shared Task – 2nd Call for Participation m

We are pleased to invite additional teams to join the AraGenre Shared Task at ArabicNLP @ EMNLP 2026. 

AraGenre is a challenging shared task on hierarchical, definition-guided Arabic genre classification, designed to evaluate how well NLP systems can generalise from limited training data to unseen genre-definition combinations.

We are delighted by the growing interest in the task, with more than 25 teams already registered. Several participants have already achieved impressive scores on the development leaderboard, making the competition increasingly challenging.  

Participants will have the opportunity to describe their approaches in a shared task system paper to appear in the final proceedings  

Can your system beat the current top-performing approaches? 

Registration: https://forms.gle/RAi7SpYoRD5RgUnP6 Codabench: https://www.codabench.org/competitions/16356 


Organisers: 

• Mo El-Haj, VinUniversity (Vietnam) / Lancaster University (UK) 

• Saad Ezzini, KFUPM (Saudi Arabia) 

• Mustafa Jarrar, HBKU (Qatar) 

• Shadi Abudalfa, KFUPM (Saudi Arabia)

Reply all
Reply to author
Forward
0 new messages