@inproceedings{hennara-etal-2026-misraj,
title = "Misraj {AI} at {AR}-{MS} {NAKBA}-{NLP} 2026: A State-of-the-Art {VLM} in {A}rabic Handwritten Text Recognition",
author = "Hennara, Khalil and
Hreden, Muhammad and
Aldallal, Zeina and
Chrouf, Sara and
AlModhayan, Safwan",
editor = "Jarrar, Mustafa and
El-Haj, Mo and
Haddad, Amal and
Atiani, Serin and
Abudalfa, Shadi and
Regier, Terry and
Rayson, Paul and
Sima{'}an, Khalil and
Mansour, Camille",
booktitle = "Proceedings of the 2nd International Workshop on Nakba Narratives as Language Resources @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.nakbanlp-1.27/",
doi = "10.63317/2uur73vrcmk5",
pages = "196--200",
abstract = "Handwritten Text Recognition (HTR) for Arabic presents unique challenges due to the script{'}s cursive nature, varying writer styles, and morphological complexity. While modern Vision-Language Models (VLMs) have significantly advanced document parsing, their direct application to highly specific cursive domains requires strategic adaptation. This paper details our submission to the Nakba OCR competition, which adapts a 3B-parameter VLM to recognize historical Arabic manuscripts. We employ a progressive training pipeline that utilizes domain-matched data augmentation to bridge the gap between standard printed Arabic OCR and historical handwritten manuscripts. Moving beyond standard decoder-only Supervised Fine-Tuning (SFT), we fine-tune the entire encoder-decoder architecture using differential learning rates. This approach, followed by a final checkpoint merge, allows the model to better resolve the fine visual details of cursive Arabic script. Our final unified model (submitted under the team name Misraj AI) establishes a new state-of-the-art (SOTA) on the Nakba dataset, achieving a Word Er- ror Rate (WER) of 0.24 and a Character Error Rate (CER) of 0.08, and officially securing first place on the leaderboard."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hennara-etal-2026-misraj">
<titleInfo>
<title>Misraj AI at AR-MS NAKBA-NLP 2026: A State-of-the-Art VLM in Arabic Handwritten Text Recognition</title>
</titleInfo>
<name type="personal">
<namePart type="given">Khalil</namePart>
<namePart type="family">Hennara</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Muhammad</namePart>
<namePart type="family">Hreden</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zeina</namePart>
<namePart type="family">Aldallal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sara</namePart>
<namePart type="family">Chrouf</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Safwan</namePart>
<namePart type="family">AlModhayan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2nd International Workshop on Nakba Narratives as Language Resources @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mustafa</namePart>
<namePart type="family">Jarrar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mo</namePart>
<namePart type="family">El-Haj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amal</namePart>
<namePart type="family">Haddad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Serin</namePart>
<namePart type="family">Atiani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shadi</namePart>
<namePart type="family">Abudalfa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Terry</namePart>
<namePart type="family">Regier</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Rayson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Khalil</namePart>
<namePart type="family">Sima’an</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Camille</namePart>
<namePart type="family">Mansour</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Handwritten Text Recognition (HTR) for Arabic presents unique challenges due to the script’s cursive nature, varying writer styles, and morphological complexity. While modern Vision-Language Models (VLMs) have significantly advanced document parsing, their direct application to highly specific cursive domains requires strategic adaptation. This paper details our submission to the Nakba OCR competition, which adapts a 3B-parameter VLM to recognize historical Arabic manuscripts. We employ a progressive training pipeline that utilizes domain-matched data augmentation to bridge the gap between standard printed Arabic OCR and historical handwritten manuscripts. Moving beyond standard decoder-only Supervised Fine-Tuning (SFT), we fine-tune the entire encoder-decoder architecture using differential learning rates. This approach, followed by a final checkpoint merge, allows the model to better resolve the fine visual details of cursive Arabic script. Our final unified model (submitted under the team name Misraj AI) establishes a new state-of-the-art (SOTA) on the Nakba dataset, achieving a Word Er- ror Rate (WER) of 0.24 and a Character Error Rate (CER) of 0.08, and officially securing first place on the leaderboard.</abstract>
<identifier type="citekey">hennara-etal-2026-misraj</identifier>
<identifier type="doi">10.63317/2uur73vrcmk5</identifier>
<location>
<url>https://aclanthology.org/2026.nakbanlp-1.27/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>196</start>
<end>200</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Misraj AI at AR-MS NAKBA-NLP 2026: A State-of-the-Art VLM in Arabic Handwritten Text Recognition
%A Hennara, Khalil
%A Hreden, Muhammad
%A Aldallal, Zeina
%A Chrouf, Sara
%A AlModhayan, Safwan
%Y Jarrar, Mustafa
%Y El-Haj, Mo
%Y Haddad, Amal
%Y Atiani, Serin
%Y Abudalfa, Shadi
%Y Regier, Terry
%Y Rayson, Paul
%Y Sima’an, Khalil
%Y Mansour, Camille
%S Proceedings of the 2nd International Workshop on Nakba Narratives as Language Resources @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F hennara-etal-2026-misraj
%X Handwritten Text Recognition (HTR) for Arabic presents unique challenges due to the script’s cursive nature, varying writer styles, and morphological complexity. While modern Vision-Language Models (VLMs) have significantly advanced document parsing, their direct application to highly specific cursive domains requires strategic adaptation. This paper details our submission to the Nakba OCR competition, which adapts a 3B-parameter VLM to recognize historical Arabic manuscripts. We employ a progressive training pipeline that utilizes domain-matched data augmentation to bridge the gap between standard printed Arabic OCR and historical handwritten manuscripts. Moving beyond standard decoder-only Supervised Fine-Tuning (SFT), we fine-tune the entire encoder-decoder architecture using differential learning rates. This approach, followed by a final checkpoint merge, allows the model to better resolve the fine visual details of cursive Arabic script. Our final unified model (submitted under the team name Misraj AI) establishes a new state-of-the-art (SOTA) on the Nakba dataset, achieving a Word Er- ror Rate (WER) of 0.24 and a Character Error Rate (CER) of 0.08, and officially securing first place on the leaderboard.
%R 10.63317/2uur73vrcmk5
%U https://aclanthology.org/2026.nakbanlp-1.27/
%U https://doi.org/10.63317/2uur73vrcmk5
%P 196-200
Markdown (Informal)
[Misraj AI at AR-MS NAKBA-NLP 2026: A State-of-the-Art VLM in Arabic Handwritten Text Recognition](https://aclanthology.org/2026.nakbanlp-1.27/) (Hennara et al., NakbaNLP 2026)
ACL