Home

The 1st Workshop on NLP for Languages Using Arabic Script

(AbjadNLP 2025)

Abu Dhabi, UAE

19 January 2025 (in person)

Submission URL: https://softconf.com/coling2025/AbjadNLP25/

Co-located with COLING 2025 Conference, Abu Dhabi, UAE (19-20 January 2025)

AbjadNLP is dedicated to advancing innovation and gaining deeper insights into Natural Language Processing (NLP) for languages that use the Arabic script. Our primary focus is on Abjad and Ajami languages that utilise the Arabic script or its variations. Traditionally associated with Semitic languages, Abjad scripts represent consonants in every syllable. In contrast, Ajami scripts denote the alphabetic use of the Arabic script in various African contexts, representing non-Arabic languages. We are interested in research on languages that fall under the Abjad or Ajami categories that use the Arabic script or any variations of it.

We invite contributions, discussions, and explorations that delve deep into the unique linguistic structures, resources, challenges, and untapped potential presented by Abjad and Ajami languages within the realm of NLP and language resources. Our goal is to create synergies among researchers by addressing the diverse phenomena and challenges inherent in these rich linguistic traditions.

The workshop is proud to highlight our connections with the Masakhane NLP community and collaborations with institutions worldwide, such as COMSATS on Urdu, and the long-standing UCREL NLP Group at Lancaster University, whose work encompasses over 20 languages worldwide, including Abjad and Ajami languages.

Note: We chose the name Abjad for simplicity, but our focus includes Abjad and other languages that have adopted the Arabic and Perso-Arabic scripts, as well as Ajami languages. We acknowledge that Sorani Kurdish, when written in Arabic script, follows an alphabet style rather than an Abjad style.

 

Workshop Description:

We welcome contributions, discussions, and explorations that thoroughly investigate the distinctive linguistic structures, resources, challenges, and untapped potential of Abjad and Ajami languages within the field of NLP and language resources. Our aim is to foster collaboration among researchers by addressing the varied phenomena and challenges inherent in these rich linguistic traditions.

Ajami languages, representing a myriad of African languages that have adopted the Arabic script, span at least 43 distinct languages, including Hausa, Fulfulde, Mandingo, Swahili, Wolof, Kanuri, and Tamazight. The combined number of speakers of these languages is estimated to exceed 200 million within Africa alone. Although Abjad has been traditionally associated with Semitic languages such as Arabic, Hebrew and Syriac, it has been adopted for writing by many other language communities as in Perso-Arabic scripts used in Persian, Urdu, Pashto, Sorani Kurdish, Azeri Turkish, Sindhi, and Uyghur, with a collective estimated speaker population exceeding 500 million. Altogether, these languages represent an approximate global aggregate of 1 billion speakers.
The adoption of the Arabic script across diverse linguistic landscapes highlights its expansive and varied application, transcending genres such as governmental correspondences, poetic compositions, religious texts, and journalistic pursuits. This widespread use underscores the imperative need to enhance digital infrastructure, tools, and resources for these under-resourced languages. Advancing such resources is crucial to nurturing linguistic diversity and resilience in both digital and print media, ensuring the preservation of linguistic heritage in the digital age.
Currently, there is an increasing interest in various NLP communities, both in academia and industry, in writing systems. However, there is a lack of initiatives focusing on the diverse phenomena and challenges of the languages using an Abjad script. The AbjadNLP workshop aims to fill this gap, fostering collaboration and innovation in this vital area of study.

 

Motivation

Languages employing an Abjad script signify a pivotal and diverse fragment of the global linguistic mosaic, traversing numerous countries and regions and embodying a considerable populace of speakers. The linguistic wealth and geographical diffusion of languages covered by AbjadNLP present a prolific environment for exploration and advancement in NLP. By channeling attention towards these languages, the realm of NLP is poised to unlock access to an expansive and varied array of linguistic constructions, subtleties, and cultural contexts, pivotal for bolstering the versatility and adaptability of NLP models and applications. The extensive spectrum of these languages not only unfolds a valuable opportunity to amplify multilingualism and multiculturalism in NLP research but also forges pathways for addressing the requisites and challenges intrinsic to a diverse and extensive speaker population.
The broad adoption of Abjad scripts transcends diverse genres, including governmental correspondences, poetic compositions, religious texts, and journalistic pursuits. The sustained use of such scripts underscores the imperative need to enhance digital infrastructure, tools, and resources that elucidate the varied writing systems inherent to under-resourced languages. Such advancement is crucial to nurturing linguistic diversity and resilience in both digital and print media, ensuring that the linguistic heritage does not diminish in the digital age.
This workshop can contribute to more inclusive and equitable progressions in NLP, accommodating a broader assortment of languages and dialects and promoting enhanced comprehension and interconnectivity amongst varied linguistic communities. The assimilation and prioritization of these linguistically affluent and diverse languages are indispensable for the comprehensive progression and the universal adaptability of NLP technologies. While our workshop primarily targets languages using an Abjad script, we recognize that many historical languages such as Aramaic , Sogdian, Parthian and Phoenician employed such a writing system. As such, we believe that our workshop can enforce links with researchers working on endangered languages as well.

We are proud to highlight our existing connections with the Masakhane NLP community (www.masakhane.io) and collaborations with institutions worldwide, such as COMSAT on Urdu (www.comsats.edu.pk), and the long-standing UCREL NLP Group at Lancaster University, whose work encompasses over 20 languages worldwide, including Abjad and Ajami languages (http://ucrel-web-dev.lancs.ac.uk/ucrelng/).

 

Team

Our team is uniquely diverse and gender-balanced, comprising individuals from a wide range of ethnic backgrounds. We represent a spectrum of languages that use the Arabic script and include researchers from both Linguistics and NLP, enriching the ever-needed collaboration between these two fields. With expertise in language technology, Unicode, NLP, resources, and multilingual text analysis, together, we aim to foster a dynamic and inclusive environment for research and collaboration in the field of NLP.

Call for papers

We invite submissions on topics that include, but are not limited to, the following:
• Enabling core technologies: morphological analysis, disambiguation, tokenisation, POS tagging, named entity detection, chunking, parsing, semantic role labelling, sentiment analysis, language modelling, etc.
• Applications: machine translation, speech recognition, speech synthesis, optical character recognition, pedagogy, assistive technologies, social media, etc.
• Resources: dictionaries, annotated data, corpus, etc.

In addition, we extend a warm invitation to researchers and stakeholders across the spectrum to contribute papers focusing on, but not limited to, the following dimensions:

  • Orthography descriptions (Constable 2002; Hosken 2003)
  • Advancements in Font Technology, Glyph Rendering, and OCR
  • Text Input Methodologies
  • Development and Utilisation of Exploitable Dictionaries
  • Enhancements in Spell-Checking Support
  • Advancements in Speech-to-Text Solutions
  • Progressive Natural Language Processing Techniques
  • BLARK – Basic Language Resource Kit descriptions for languages using Abjad or Ajami
  • Insights and Experiences Utilising Data Supplied by the Unicode Hosted Common Locale Data Repository in Abjad or Ajami.
  • Morphological and syntactical challenges in Abjad or Ajami Orthographies.
  • Development of open access corpora in Abjad or Ajami.
  • Text processing and transliteration challenges and solutions for languages using Abjad or Ajami.
  • Cultural and sociolinguistic considerations in NLP applications for Abjad or Ajami.
  • Languages resources and NLP tools for Abjad or Ajami.

 

Summary of the Call:

We welcome submissions of papers centred around the Abjad and Ajami theme, focusing on supporting NLP language resources for non-Arabic languages utilising Arabic script. We encourage submissions that span a spectrum from theoretical investigations to practical applications, aiming to underscore the distinctive challenges, solutions, and insights that languages using Ajami and Abjad scripts introduce to the field of NLP.

For the submission format and guidelines, we follow the COLING 2025 standards. Authors are encouraged to thoroughly review and adhere to the COLING 2025 submission guidelines and author kit, which can be found at: https://coling2025.org/calls/submission_guidlines/.

If authors are describing an orthography, we request that they include the points recommended in (Hosken 2003 https://scripts.sil.org/WP-Encoding). For continuity across the workshop and greater impact across industry applications, authors should consider terminological (orthography, script, writing system, etc.) differences presented by Constable (2002) https://www.sil.org/resources/publications/entry/7853. The model presented by Constable is the current Unicode model.  

Please ensure that all submissions strictly conform to these standards to streamline the review process and maintain uniformity across all contributions. Both long papers (up to 8 pages) and short papers (up to 4 pages) are welcome. All submissions will undergo a rigorous peer-review process, emphasizing originality, relevance, and clarity.

 

Submissions may be of two types:

  1. Long papers – up to eight (8) pages maximum, presenting substantial, original, completed, and unpublished work.
  2. Short papers – up to four (4) pages, describing a small focused contribution, negative results, system demonstrations, etc.

Submission URL: https://softconf.com/coling2025/AbjadNLP25/

Submission Guidelines: https://coling2025.org/calls/submission_guidlines/

Provisional Key Dates:

  • 1st Call for Papers Announcement: 16 July 2024
  • 2nd Call for Papers Announcement: 16 August 2024
  • Paper Submission Deadline: 15 November 2024 2 December 2024 (extended)
  • Notification of Paper Acceptance: 3 December 2024
  • Camera-ready Paper Deadline: 10 December 2024
  • Workshop Date: 19  January 2024 (in person)

 

Anti-Harassment Policy:

The workshop supports the COLING anti-harassment policy https://coling2022.org/policy

Organising Committee:

General Chair:

  • Dr. Mo El-Haj, Senior Lecturer at Lancaster University, is a Natural Language Processing expert with a focus on Arabic and under-resourced languages. He founded the FNP workshop series in 2018 and has organised workshops at top NLP conferences including LREC and COLING. http://elhaj.uk/

Programme Chairs:

Review Committee:

Publication Chair:

  • Dr Sina Ahmadi. Postdoctoral researcher at University of Zurich focusing on leveraging language technology to assist languages with constrained digital resources, with an emphasis on adapting current natural language processing approaches and existing resources for less-resourced languages. https://sinaahmadi.github.io/

Publicity Chairs:

Advisory Committee:

  • Prof. Ruslan Mitkov, Professor of Computing and Communications at Lancaster University, actively working on different research topics from the areas of Natural Language Processing (NLP), Computational Linguistics and Translation Technology.https://wp.lancs.ac.uk/mitkov/
  • Prof. Paul Rayson, Director of UCREL research centre at Lancaster University, specialises in semantic-based NLP across 20 languages, including Urdu and Arabic. With 25 years of experience, he excels in noisy language environments like financial disclosures and has organised various conferences and workshops. https://www.lancaster.ac.uk/staff/rayson/

 

List of Accepted papers:

Title Authors
Developing an Informal-Formal Persian Corpus: Highlighting the Differences between Two Writing Styles Vahide Tajalli, Mehrnoush Shamsfard and Fateme Kalantari
DadmaTools V2: an Adapter-Based Natural Language Processing Toolkit for the Persian Language sadegh jafari, Farhan Farsi, Navid Ebrahimi, Mohamad Bagher Sajadi and Sauleh Eetemadi
Towards Unified Processing of Perso-Arabic Scripts for ASR Srihari Bandarupalli, Bhavana Akkiraju, Sri Charan Devarakonda, Harinie Sivaramasethu, Vamshiraghusimha Narasinga and Anil Vuppala
Evaluating Large Language Models on Health-Related Claims Across Arabic Dialects Abdulsalam obaid Alharbi, Abdullah Alsuhaibani, Abdulrahman Abdullah Alalawi, Usman Naseem, Shoaib Jameel, salil kanhere and Imran Razzak
Can LLMs Verify Arabic Claims? Evaluating the Arabic Fact-Checking Abilities of Multilingual LLMs Ayushman Gupta, Aryan Singhal, Thomas Law, Veekshith Rao, Evan Duan and Ryan Luo Li
Can LLMs Translate Cultural Nuance in Dialects? A Case Study on Lebanese Arabic. silvana yakhni and Ali Chehab
Boosting Sentiment Analysis in Persian through a GAN-Based Synthetic Data Augmentation Method Masoumeh Mohammadi, Mohammad Ruhul Amin and Shadi Tavakoli
In-Depth Analysis of Arabic-Origin Words in the Turkish Morpholex Mounes Zaval, Abdullah İhsanoğlu, Asım Ersoy and Olcay Taner Yıldız
A Derivational ChainBank for Modern Standard Arabic Reham Marzouk, sondos krouna and Nizar Habash
Sentiment Analysis of Arabic Tweets Using Large Language Models Pankaj Dadure, Ananya Dixit, Kunal Tewatia, Nandini Paliwal and Anshika Malla
Evaluation of Large Language Models on Arabic Punctuation Prediction Asma Ali Al Wazrah, Afrah Altamimi, Hawra Aljasim, Waad Alshammari, Rawan Al-Matham, Omar Elnashar, Mohamed Amin and Abdulrahman AlOsaimy
Evaluating RAG Pipelines for Arabic Lexical Information Retrieval: A Comparative Study of Embedding and Generation Models Raghad Al-Rasheed, Abdullah Al Muaddi, Hawra Aljasim, Rawan Al-Matham, Muneera Alhoshan, Asma Al Wazrah and Abdulrahman AlOsaimy
Automated Generation of Arabic Verb Conjugations with Multilingual Urdu Translation: An NLP Approach Haq Nawaz, Manal Elobaid, Ali Al-Laith and Saif Ullah
Psychological Health Chatbot, Detecting and Assisting Patients in their Path to Recovery sadegh jafari, Mohammad Erfan Zare, Amireza Vishte, Mirzae Melike, Zahra Amiri, Sima Mohammadparast and Sauleh Eetemadi
MultiProp Framework: Ensemble Models for Enhanced Cross-Lingual Propaganda Detection in Social Media and News using Data Augmentation, Text Segmentation, and Meta-Learning Farizeh Aldabbas, Shaina Ashraf, Rafet Sifa and Lucie Flek
The Best of Both Worlds: Exploring Wolofal in the Context of NLP Ngoc Tan Le, Ali Mijiyawa, Abdoulahat Leye and Fatiha Sadat

 

Programme Committee* 

  • Abdoulaye Diallo. Fula & Wolof. Independent Researcher
  • Ahmed Abdelali. Arabic/Multilingual NLP. Qatar Computing Research Institute (QCRI), Qatar.
  • Ahmed AbuRa’ed. Arabic. UBC. Canada.
  • Antonio Moreno Sandoval. Low Resourced Languages. UAM. Spain
  • Azizud Din. Pashto. University Malaysia Sarawak. Malaysia
  • Behnam Sabeti. Persian. Miras Technologies International. Iran
  • Chenggang Mi. Uyghur. Xinjiang Technical Institute. China
  • Clement Oyeleke. Yoruba. University of Ibadan. Nigeria
  • Daniel Whitenack. Kimbundu, Fulfulde, Pular. SIL International. USA
  • Derguene Mbaye. Wolof. Baamtu. Senegal
  • Djamel Mostefa. Pashto. ELDA, France.
  • Doaa Samy. Arabic. Cairo University, Egypt and LLI-UAM. Spain
  • Elias W BA. Fula and Wolof. Baamtu. Senegal
  • Frederick Apina. Swahili. Parrot.AI. Tanzania
  • George Giannakopoulos. Multilingual NLP. SKEL Lab – NCSR Demokritos. Greece
  • Haithem Afli. Arabic/Multilingual NLP. Dublin City University, Ireland.
  • Hazem Hajj. Arabic/Multilingual NLP. American University of Beirut, Lebanon.
  • Houda Bouamor. Arabic/Multilingual NLP. CMU. Qatar
  • Ignatius Ezeani. Igbo, African Languages NLP. Lancaster University, UK.
  • Imed Zitouni. Arabic/Multilingual NLP. Microsoft Research, USA.
  • Karim Bouzoubaa. Arabic/Multilingual NLP. Mohamed Vth University, Morocco.
  • Mariam Masoud. Sorani Kurdish. National University of Ireland Galway. Ireland
  • Lei Wang. Uyghur. Xinjiang Technical Institute. China
  • Marina Litvak. Hebrew and Arabic. Sami Shamoon College of Engineering, Israel
  • Mo El-Haj. Arabic/Multilingual and Low resourced Languages. Lancaster University, UK
  • Muhammad Sharjeel. Urdu. COMSATS University Islamabad, Pakistan.
  • Omid Momenzadeh. Persian. Miras Technologies International. Iran
  • Paul Rayson. Multilingual and Low resourced Languages. Lancaster University, UK
  • Preni Golazizian. Persian. Miras Technologies International. Iran
  • Rao Muhammad Adeel Nawab.  Urdu. COMSATS University Islamabad, Pakistan.
  • Reza Fahmi. Persian. Miras Technologies International. Iran
  • Samuel Olanrewaju. Yoruba, Yagba and Basa. University of Ibadan. Nigeria
  • Scott Piao. Multilingual and Low resourced Languages. Lancaster University, UK
  • Seyed Arad Ashrafi Asli. Persian. Miras Technologies International. Iran
  • Shervin Malmasi. Sorani Kurdish. Macquarie University. Australia
  • Sina Ahmadi. Sorani Kurdish. National University of Ireland Galway. Ireland
  • Sokhar Samb. Wolof. ML & NLP. Senegal
  • Tonghai Jiang. Uyghur. Xinjiang Technical Institute. China
  • Waziri Shebogholo. Swahili. Parrot.AI. Tanzania
  • Wole Akin. IsiXhosa, Yorùbá, Hausa, and Igbo. University of Johannesburg. South Africa
  • Xi Zhou. Uyghur. Xinjiang Technical Institute. China
  • Yating Yang. Uyghur. Xinjiang Technical Institute. China
  • Zahra Majdabadi. Persian. Miras Technologies International. Iran

 

*We are in the process of forming a linguistically diverse program committee who are experts in languages that use Arabic Script (Abjad and Ajami), with the majority of the list already confirmed to serve as reviewers. As soon as we gain access to SoftConf, we will extend invitations to the remaining committee (if you see your name on the list and want it removed, please contact any of the organisers). If your name appears in this list and you want it removed, please contact us as soon as possible and we’ll make sure it’s removed. Thanks