@inproceedings{ibrahim-etal-2026-alexis,
title = "{ALEXIS} at {A}lexandria{X}-2026 Subtask 2: Capacity Discovery and Calibrated Preference Optimization for Cross-Dialect {A}rabic {MT}",
author = "Ibrahim, Momen and
Hassib, Mariam E. and
Mahmoud Salem, Ahmed and
Torki, Marwan",
editor = "Abdul-Mageed, Muhammad and
Elsayed, Tamer and
Bouamor, Houda and
El-Beltagy, Samhaa and
Al-Khalifa, Hend and
Abdelali, Ahmed and
Elmadany, AbdelRahim and
El Mekki, Abdellah and
Darwish, Kareem and
Alkhereyf, Sakhar and
Luqman, Hamzah and
Khalifa, Salam and
Magdy, Samar and
Alyafeai, Zaid and
Hamed, Injy and
Zaghouani, Wajdi and
Zitouni, Imed",
booktitle = "Proceedings of The Fourth {A}rabic Natural Language Processing Conference: Shared Tasks",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.arabicnlp-sharedtasks.104/",
pages = "834--841",
ISBN = "979-8-89176-497-2",
abstract = "Cross-dialect Arabic machine translation asks for translation between Modern Standard Arabic and regional dialects with almost no parallel text. We present a system for Subtask 2 of AlexandriaX-2026, built on two findings about where the gains come from. First, data was not the binding constraint: once our mined-and-synthetic recipe had decelerated to $+1.8\%$ relative spBLEU per round, a single increase in LoRA capacity beat three further data-tuning rounds combined ($+8.9\%$). Second, planned epoch budgets were consistently too short, and training past them was the largest single lever at every stage. We fine-tune NLLB-200-distilled-600M on a three-tier corpus of real, mined and synthetic pairs, then refine it with Direct Preference Optimization. Ablating three reward designs isolates a third finding: rewarding agreement with self-distilled references regresses real quality despite healthy training dynamics. Restricting the reward to references we trust, with a reference-free substitute for the two zero-resource dialects, yields a real gain that a further iteration gives back. Every configuration was scored on the official leaderboard. Our submission reached spBLEU 25.84 and chrF++ 39.02, ranking 2nd of 4 teams; post-evaluation experimentation reached 28.04."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ibrahim-etal-2026-alexis">
<titleInfo>
<title>ALEXIS at AlexandriaX-2026 Subtask 2: Capacity Discovery and Calibrated Preference Optimization for Cross-Dialect Arabic MT</title>
</titleInfo>
<name type="personal">
<namePart type="given">Momen</namePart>
<namePart type="family">Ibrahim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mariam</namePart>
<namePart type="given">E</namePart>
<namePart type="family">Hassib</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Mahmoud Salem</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marwan</namePart>
<namePart type="family">Torki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of The Fourth Arabic Natural Language Processing Conference: Shared Tasks</title>
</titleInfo>
<name type="personal">
<namePart type="given">Muhammad</namePart>
<namePart type="family">Abdul-Mageed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tamer</namePart>
<namePart type="family">Elsayed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samhaa</namePart>
<namePart type="family">El-Beltagy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hend</namePart>
<namePart type="family">Al-Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Abdelali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">AbdelRahim</namePart>
<namePart type="family">Elmadany</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abdellah</namePart>
<namePart type="family">El Mekki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kareem</namePart>
<namePart type="family">Darwish</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sakhar</namePart>
<namePart type="family">Alkhereyf</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hamzah</namePart>
<namePart type="family">Luqman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Salam</namePart>
<namePart type="family">Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samar</namePart>
<namePart type="family">Magdy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zaid</namePart>
<namePart type="family">Alyafeai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Injy</namePart>
<namePart type="family">Hamed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wajdi</namePart>
<namePart type="family">Zaghouani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Imed</namePart>
<namePart type="family">Zitouni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-497-2</identifier>
</relatedItem>
<abstract>Cross-dialect Arabic machine translation asks for translation between Modern Standard Arabic and regional dialects with almost no parallel text. We present a system for Subtask 2 of AlexandriaX-2026, built on two findings about where the gains come from. First, data was not the binding constraint: once our mined-and-synthetic recipe had decelerated to +1.8% relative spBLEU per round, a single increase in LoRA capacity beat three further data-tuning rounds combined (+8.9%). Second, planned epoch budgets were consistently too short, and training past them was the largest single lever at every stage. We fine-tune NLLB-200-distilled-600M on a three-tier corpus of real, mined and synthetic pairs, then refine it with Direct Preference Optimization. Ablating three reward designs isolates a third finding: rewarding agreement with self-distilled references regresses real quality despite healthy training dynamics. Restricting the reward to references we trust, with a reference-free substitute for the two zero-resource dialects, yields a real gain that a further iteration gives back. Every configuration was scored on the official leaderboard. Our submission reached spBLEU 25.84 and chrF++ 39.02, ranking 2nd of 4 teams; post-evaluation experimentation reached 28.04.</abstract>
<identifier type="citekey">ibrahim-etal-2026-alexis</identifier>
<location>
<url>https://aclanthology.org/2026.arabicnlp-sharedtasks.104/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>834</start>
<end>841</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T ALEXIS at AlexandriaX-2026 Subtask 2: Capacity Discovery and Calibrated Preference Optimization for Cross-Dialect Arabic MT
%A Ibrahim, Momen
%A Hassib, Mariam E.
%A Mahmoud Salem, Ahmed
%A Torki, Marwan
%Y Abdul-Mageed, Muhammad
%Y Elsayed, Tamer
%Y Bouamor, Houda
%Y El-Beltagy, Samhaa
%Y Al-Khalifa, Hend
%Y Abdelali, Ahmed
%Y Elmadany, AbdelRahim
%Y El Mekki, Abdellah
%Y Darwish, Kareem
%Y Alkhereyf, Sakhar
%Y Luqman, Hamzah
%Y Khalifa, Salam
%Y Magdy, Samar
%Y Alyafeai, Zaid
%Y Hamed, Injy
%Y Zaghouani, Wajdi
%Y Zitouni, Imed
%S Proceedings of The Fourth Arabic Natural Language Processing Conference: Shared Tasks
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-497-2
%F ibrahim-etal-2026-alexis
%X Cross-dialect Arabic machine translation asks for translation between Modern Standard Arabic and regional dialects with almost no parallel text. We present a system for Subtask 2 of AlexandriaX-2026, built on two findings about where the gains come from. First, data was not the binding constraint: once our mined-and-synthetic recipe had decelerated to +1.8% relative spBLEU per round, a single increase in LoRA capacity beat three further data-tuning rounds combined (+8.9%). Second, planned epoch budgets were consistently too short, and training past them was the largest single lever at every stage. We fine-tune NLLB-200-distilled-600M on a three-tier corpus of real, mined and synthetic pairs, then refine it with Direct Preference Optimization. Ablating three reward designs isolates a third finding: rewarding agreement with self-distilled references regresses real quality despite healthy training dynamics. Restricting the reward to references we trust, with a reference-free substitute for the two zero-resource dialects, yields a real gain that a further iteration gives back. Every configuration was scored on the official leaderboard. Our submission reached spBLEU 25.84 and chrF++ 39.02, ranking 2nd of 4 teams; post-evaluation experimentation reached 28.04.
%U https://aclanthology.org/2026.arabicnlp-sharedtasks.104/
%P 834-841
Markdown (Informal)
[ALEXIS at AlexandriaX-2026 Subtask 2: Capacity Discovery and Calibrated Preference Optimization for Cross-Dialect Arabic MT](https://aclanthology.org/2026.arabicnlp-sharedtasks.104/) (Ibrahim et al., ArabicNLP 2026)
ACL