The Second Arabic NLP School

About the School

The Second Arabic NLP School 2026 is a one-day event co-located with EACL 2026 in Rabat, Morocco, on Tuesday, March 24, 2026.  The first edition of the Arabic NLP School (Arabic NLP Winter School 2025) was co-located with COLING 2025 in Abu Dhabi, UAE (January 18-19, 2025). 100 people participated in the First Arabic NLP Winter School.

The 2026 edition of the School introduces a redesigned, highly practical format aimed at developing end-to-end research skills in NLP.  The School will guide participants through the full research lifecycle: from conceiving an idea, formulating research questions, conducting responsible and thorough related-work reviews, designing meaningful experiments, and curating/annotating data, all the way to preparing submissions, producing camera-ready papers, and effectively presenting research at conferences.   While these skills are central to any NLP research effort,  the School focuses specifically on Arabic language technologies, Arab cultures, and the unique needs of emerging researchers from the Arab world.  

The program is an intensive, hands-on experience combining expert panels, ethics-focused discussions, and collaborative teamwork to develop a complete, well-founded research proposal in Arabic NLP.   The program seeks to empower the next generation of scholars and practitioners who will shape the future of Arabic NLP.

We anticipate a wide range of participants, including researchers, undergraduate and graduate students, PhD candidates, and industry and government practitioners with an interest in Arabic NLP.

Arabic NLP School 2026 is an application-based program offered free of charge to applicants who are accepted (thanks to our great ✨sponsors✨!). Application details are provided below.

The Winning Teams

🥇 TurathBench 

• Project: WarshRec: A Benchmark for Quranic Tajweed Recitation Error Detection and Correction

• Team: Noureddine Khaous, Ismael, Amal El Mahraoui, Arif Hiba, Outhmane Marmouzi, and Fatmazzahra Elhoubri

• Mentors: Kareem Darwish and Dris Namly

🥈 PolyArab 

• Project: Distinguishing Dialect from Disorder: Robust Clinical NLP Biomarkers for Arabic Speech

• Team: Ghofrane Merhbene, Khalid Abid, Zakaria El-Aoufi, Nouhaila Houssa, and Zakaria Zaouak

• Mentors: Hakim Hafidi and Khalil Mrini

🥉 The Sentimentals 

• Project: Thinking Emotionally in Arabic

• Team: Aya Cherqi, Safae Bouhaddou, Fatima-zahra Aazi, Ikrame Kiyadi, Maida Aizaz, Mourad El Asmai, Nouhaila Rabii, and Rabia Rachidi

• Mentors: Youness Moukafih and Haithem Afli

Venue

The School will take place at the EACL 2026 conference venue: Palais des Congrès Rabat Bouregreg (Google Maps).

Proposed Projects

(1) Dialect Modeling & Cross-Variety NLP

• Fuṣḥā'izi: Fusḥa wa Fuṣḥā: From dialectal wording to semi-MSA

  • Dialect - Team 1: Rayyan Merchant (Lead), Amir Ejmail, Mouaad Errami, Walid Kerroumi, Ismail El Bazi, Hanae Baraka

  • Mentors: Walid Magdy and Salam Khalifa

• Natakallam: Enhancing Arabic Dialect Understanding with Lexicons and Synthetic Data

  • Dialect - Team 2: Hiba Bouhnin (Lead), Nathaniel Robinson, Yassine Farah, Khalil Mellouk, Yassine Bouras, Youssef Mahdoubi, Rebbah Yahya

  • Mentors: Salam Khalifa and Walid Magdy

• PolyArab: How Well Do LLMs Really Understand Arabic Dialects? A Benchmarking Study

  • Dialect - Team 4: Ghofrane Merhbene (Lead), Khalid Abid, Zakaria El-Aoufi, Nouhaila Houssa, and Zakaria Zaouak.

  • Mentors: Hakim Hafidi and Khalil Mrini

(2) Arabic Data Resources & Benchmarking

• Tibyan: Corpus-Grounded Retrieval-Augmented Generation for Reliable LLM-Based Arabic Reading Assessment

  • Data Resources - Team 1: Noura Ogbi (Lead), Samir Abdaljalil, Hassan Oukhouya, and Hafssa Ziyati.

  • Mentors: Hamdy Mubarak and Go Inoue

• ArabicRAG-Eval: A Faithfulness Benchmark for Arabic Retrieval-Augmented Generation Systems

  • Data Resources - Team 2: Zaineb Rahmani (Lead), Abderrahmane Jouilili, Abdellah Hasnaoui, Moumni Mohammed, and Wissal Saib.

  • Mentors: Go Inoue and Hamdy Mubarak

• TurathBench: MaktaBench: Advancing Arabic NLP through Benchmarking & Dataset Documentation

  • Data Resources - Team 3: Ismail El Jamiy (Lead), Hiba Arif, Larbi Boulaarab, Noureddine Khaous, Fatima Ezzahra El Houbri, Amal El Mahraoui, Majda Essaadi, and Outmane Marmouzi.

  • Mentors: Kareem Darwish and Driss Namly

• TeamWise-QAMAR AI: QMHallucinate: A Benchmark for Detecting Hallucinations in LLM-Based Morphological Analysis of Quranic Arabic

  • Data Resources - Team 4: Sara Faqihi (Lead), Firdaous Ait Mohamed, Adnane El Amrani, and Omar Momen

  • Mentors: Driss Namly and Kareem Darwish

(3) Domain-Specific Arabic NLP

• NurNLP: Improving Quranic Question Answering and Benchmarking: Towards Hallucination Detection, Complex Query Handling, and Enhanced Retrieval

  • Domain-Specific - Team 1: Youssef Amzoug (Lead), Hiba Benkaddour, Oumayma Elbiach, Anas Hajbi, and Sana Khayou.

  • Mentors: Bashar Alhafni and Ehsaneddin Asgari

• DarijaMed: A Patient-Centered Assistant for Explaining French Medical Reports in Moroccan Darija

  • Domain-Specific - Team 2: Saad Frihi (Lead), Chaima Ben Jaafar, Khaoula Draoui, Wiam Makboul, and Abdelillah Omari.

  • Mentors: Ehsaneddin Asgari and Bashar Alhafni

• AraFinMerge: Accelerating large language models development for financial Arabic text classification through model merging methods.

  • Domain-Specific - Team 3: Mohamed Khenchouch (Lead), Mohamed El Amraoui, Yassine Mountassir Idrissi, and Mohammed Lahlou

  • Mentors: Lamia Ben Hiba and Tamer Elsayed

• Al-Bayyinah: AL-QADI: An LLM-as-a-Judge Framework for Hallucination Detection and Abstention in Arabic Islamic Content

  • Domain-Specific - Team 4: Amina El Ganadi (Lead), Fatima-ezzahra Darfaoui, Rokaya El Gounidi, Mouad Hakam, and Sanjeev Kumar.

  • Mentors: Tamer Elsayed and Lamia Ben Hiba

(4) Social Good and Arabic NLP

• The Sentimentals: Challenges and Opportunities in MSA Emotion Detection: insights from large language models

  • Social good - Team 1: Rabia Rachidi (Lead), Fatima-Zahra Aazi, Maida Aizaz, Aya Cherqi, Mourad El Asmai, Ikrame Kiyadi, and Nouhaila Rabii.

  • Mentors: Haithem Afli and Youness Moukafih

• Lisan Lab: Dialect-Aware Counter-Speech Generation for Reducing Toxicity in Arabic Online Discourse.

  • Social Good - Team 3: Abed Qaddoumi (Lead), Chaimae Bouhatouss, Fatima Ezzahra Chourak, and Ouail Laamiri.

  • Mentors: Kawtar Younsi Dahbi  and Wajdi Zaghouani"

• AraFair: A Multi-Dimensional Gender Bias Evaluation Framework for Arabic Large Language Models

  • Social Good - Team 4: Roaa Abdelmagid (Lead), Chaimae Abouzahir, Hammam Akrami, Rihab Atbir, Zakaria Baannou, Abdelmonim Benbouchta, Marème Diop, and Zakaria Mourid.

  • Mentors: Wajdi Zaghouani and Kawtar Younsi Dahbi

Participation and Structure

• Capacity: The school will host 120 in-person student participants. Attendance is not hybrid; all participants must be physically present.

• Selection: Participants will be chosen through an application process. Mentors, drawn from respected members of the Arabic NLP community, will take part in reviewing applications.

• Team-Based Research: Attendees will be organized into 16 teams of 6-8, each guided by a dedicated primary mentor and secondary mentor.

• Workspace: Round-table seating will accommodate teams of nine (mentor + eight team members) to encourage active collaboration and co-creation.

The goal of the team structure is to ensure that every participant gains hands-on experience in collaborative research planning and communication.

Outcomes and Impact

• Build deep, practical research skills among emerging scholars in Arabic NLP

• Cultivate long-term mentorship relationships and professional networks

• Foster collaborative projects rooted in regional linguistic and cultural needs

• Equip participants with the tools to become future leaders and contributors to SIGARAB and the broader NLP field

The school is designed to be an intensive, community-driven experience that accelerates participants’ readiness to engage in high-impact Arabic NLP research.

Organizers

• Nizar Habash, CAMeL Lab, New York University Abu Dhabi

• Houda Bouamor, Carnegie Mellon University in Qatar

Application Process (CLOSED)

• 🚀🚀🚀 The application form for The Arabic NLP School is provided 🌟🌟🌟here🌟🌟🌟.

• Important Dates

  • Application Deadline: January 31, 2026

  • Acceptance Notification:  February 9, 2026

  • Arabic NLP School: March 24, 2026.

• We reiceived over 300 applications! We selected 120 students to be part of the Arabic NLP School.

Platinum Sponsors

Gold Sponsor

Bronze Sponsor

Supporting Organizations