The Fourth
Arabic Natural Language Processing Conference
(ArabicNLP 2026)
https://arabicnlp2026.sigarab.org
Co-located with EMNLP 2026, Budapest, Hungary, October 24-29, 2026
Embedded Files
The Fourth
News
News
March 15, 2026
June 16, 2026
ArabicNLP 2026 is going to be co-located with EMNLP 2026.
ArabicNLP2026 will be running for two days: Wednesday, October 28 and Thursday, October 29
Important Dates
Important Dates
June 18, 2026: Abstract submission for conference papers due date
June 25, 2026: Direct conference paper submission due date
July 16, 2026: Reviews submission deadline
July 26 - August 2, 2026: Rebuttal period
August 7, 2026: ARR commitment deadline
August 7, 2026: Metareview deadline
August 20, 2026: Notification of acceptance
September 17, 2026: Camera-ready papers due
October 28-29, 2026: ArabicNLP conference
All deadlines are 11:59 pm UTC -12h (“Anywhere on Earth”).
Conference Description
Conference Description
The Fourth Arabic Natural Language Processing Conference (ArabicNLP 2026) is organized by the ACL Special Interest Group on Arabic NLP (SIGARAB). ArabicNLP 2025 follows the First ArabicNLP conference (in 2023 with EMNLP), the Second ArabicNLP conference (in 2024 with ACL), the third ArabicNLP conference (in 2025 with EMNLP), and seven previous workshop editions (as WANLP), which have been very successful drawing in a large active participation in various capacities (See ArabicNLP Scholar).
The research focus of ArabicNLP is, naturally, Arabic, a collection of language varieties, from Classical to Modern Standard Arabic (MSA), and including many living and historical Arabic dialects. Arabic poses many challenges for the field of computational linguistics, including rich morphology, orthographic ambiguity as well as the wide variety of understudied dialects.
Arabic is one of the official languages of the United Nations and the native tongue of over 473 million native speakers living in a region of geopolitical significance. All these factors combined make Arabic an attractive object of study from a computational perspective. The fast growth of Arabic on the Internet over the past few years and the vibrant use of Arabic dialects on social media, including issues of global relevance, lend the proposed conference particular and timely significance. The opportunities that are made possible by working on Arabic, both standard and dialects, cannot be underestimated in their consequence on the Arab world, the Mediterranean region, and the rest of the world, let alone on the advancement of computational linguistics methods and techniques to address significant inherent linguistic complexities.
We invite submissions on topics that include but are not limited to, the following:
Core Arabic NLP: methods for MSA and/or dialects, including segmentation/tokenization, morphological and morphosyntactic analysis/disambiguation, POS tagging, named entity recognition, chunking, parsing, semantic role labeling, text classification (e.g., sentiment analysis, offensiveness/hate speech), etc.
Machine learning for Arabic: theory, modeling, and systems for generative and discriminative approaches, including domain adaptation, transfer learning, multilingual/multidialectal learning, and evaluation methodology.
Arabic Large Language Models: pretraining, instruction tuning, alignment, benchmarking, robustness, fairness, and culturally grounded evaluation, etc.
Applications: machine translation, summarization, dialogue, creative generation/ analysis, ASR and TTS, OCR, educational and assistive technologies, and social media analytics.
Resources and infrastructure: corpora, lexicons/dictionaries, annotations and guidelines, benchmarks, and tools to support Arabic NLP research and deployment.
Submissions must have a clear focus on specific issues pertaining to the Arabic language whether it is standard Arabic, dialectal, or mixed. Papers on other languages sharing problems faced by Arabic NLP researchers such as Semitic languages or languages using Arabic script are welcome. Papers on generalized efforts using Arabic resources but targeting other languages are also welcome. Descriptions of commercial systems are welcome, but authors should be willing to discuss the details of their work. We accept long, short, and demo paper submissions.
Tagline: ArabicNLP 2026 presents the latest advancements in Arabic natural language processing research, resources, techniques, and applications.
Mona T. Diab is Director of the Language Technologies Institute and a Full Professor in the School of Computer Science at Carnegie Mellon University. An ACL Fellow, she has devoted more than 13 years of her career to Arabic NLP and received the 2023 King Salman Global Academy for Arabic Language prize for advancing Arabic language technologies. She is a founding member of SIGARAB, a former President of SIGDAT and SIGLEX, and a cofounder of theSEM conference and the CADIM Arabic Dialect Group. Her career spans academia and industry, including leadership roles at Meta and Amazon AWS. Her research connects multilingual NLP, cultural and social dimensions of meaning, and responsible AI. She advocates for a science of language, meaning, and interaction, with Arabic helping shape foundational questions about intelligence, understanding, and trustworthy AI.
Invited Speakers
Invited Speakers
Professor Mona Diab, Director of the Language Technologies Institute and a Full Professor in the School of Computer Science at Carnegie Mellon University
Title: From NLP for Arabic to NLP Through Arabic. What Must AI Understand Before We Trust It to Act?
Abstract: What does it mean for an AI system to understand a request, recognize an intention, or know when it has misunderstood? These questions reach into the essence of intelligence—and acquire urgency when systems act as autonomous agents. An instruction can be followed literally while its purpose is missed. A fluent answer can conceal uncertainty. Understanding a harmful request perfectly can itself enable harm. This keynote argues that the science of language, meaning, understanding, and interaction should fundamentally reorient NLP’s research agenda and inform how we assess AI’s capabilities and risks. Arabic offers a particularly productive setting for developing that science. Rich morphology, unwritten vowels, dialect diversity, diglossia, script variation, and code-switching converge within Arabic communicative encounters shaped by social relationships and historical knowledge. I introduce the Arabic Conjunction Hypothesis to investigate how these dimensions jointly shape interpretation and communicative action. Arabic’s connections with Persian, Urdu, and Turkic traditions extend the inquiry across languages, scripts, and communities. Arabic’s scientific value lies in this productive conjunction, without requiring a claim of exceptionalism. The resulting questions connect directly to debates about alignment and control: whose intentions should a system follow, what evidence supports its interpretation, and when should it ask, defer, challenge, or revise? Understanding is necessary for many forms of trustworthy assistance, but it does not establish benevolent goals or legitimate authority. A science of meaning and interaction can help make those distinctions explicit and testable. Through this lens, annotation, evaluation, and dialogue become instruments for discovering how understanding works. The shift from NLP for Arabic to NLP through Arabic makes Arabic a source of insights that can reshape the field’s questions, methods, and foundations. What would we build—and what would we measure—if understanding in interaction became NLP’s organizing purpose? Pursuing that question also challenges how we define and evaluate intelligence in AI more broadly. If we fear what AI might do, we must examine what it understands, whose purposes it serves, and what warrants trusting it to act.
Conference Size / Workshop Size / Prior Events
Estimated Number of Attendees: 75-100 in-person and 50-75 virtually. In ArabicNLP 2025, the conference received 170 people registrations.
We also managed to secure financial sponsorships to support virtual student attendance in the past.
Previous Editions: Below are the details of the previous editions leading to this proposal.
3rd ArabicNLP @ EMNLP 2025, Suzhou, China: Main conference: 95 submissions, 39 accepted (41%). Shared tasks: 11 overview papers, 127 system papers.
2nd ArabicNLP @ ACL 2024, Thailand: Main conference: 68 submissions, 31 accepted (42.6%). Shared tasks: 8 overview papers, 71 system papers.
1st ArabicNLP @ EMNLP 2023, Singapore: Main conference: 80 submissions, 38 accepted (47%). Shared tasks: 5 overview papers, 48 system papers.
7th WANLP @ EMNLP 2022, Abu Dhabi, UAE: 68 submissions, 36 accepted (52%), and 28 Shared Task system descriptions.
6th WANLP @ EACL 2021, Online: 43 submissions, 29 accepted (67%), and 24 Shared Task system descriptions.
5th WANLP @ COLING 2020, Online: 42 submissions, 3 withdrawn, 20 accepted (51%), and 15 Shared Task system descriptions.
4th WANLP @ ACL 2019 in Florence, Italy: 45 submissions; 22 were accepted (49%), and 17 Shared Task system descriptions.
3rd WANLP @ EACL 2017 in Valencia, Spain: 47 submissions; 22 accepted (47%).
2nd WANLP @ ACL 2015 in Beijing, China: 23 submissions; 15 accepted (65%), and 8 Shared Task system descriptions.
1st WANLP @ EMNLP 2014 in Doha, Qatar: 40 submissions; 20 accepted (50%), and 9 Shared Task system descriptions.
Virtual Mode: We expect ArabicNLP 2026 to be a hybrid event. There will be a mix of in-person and hybrid presenters for talks and posters. We will ask the speakers to record their presentations ahead of time and share them with the organizers.
Shared Tasks
https://arabicnlp2026.sigarab.org/shared-tasks
Organizers
Google Sites
Report abuse