@inproceedings{alkhder-etal-2026-monolingual,
title = "A Monolingual {L}o{RA}-Tuned {Q}wen2.5-3{B} System for {A}rabic Context-Based Question Answering",
author = "Alkhder, Hasan and
Hanini, Maria and
Hocine, Imane and
Pasa, Maher and
Najjar, Amro",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Eleventh Conference on Machine Translation",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.wmt-main.125/",
pages = "1899--1905",
ISBN = "979-8-89176-474-3",
abstract = "In this paper, we describe our submission to the WMT26 Multilingual Instruction Shared Task (MIST), Sub-task 1 (context-based question answering). We fine-tune Qwen2.5-3B-Instruct with a LoRA adapter on 3,112 extractive-QA examples drawn from the answerable subset of TyDi QA as released in the wmt26-mist-sample data. We submit outputs for the monolingual Arabic (question and context both in Arabic) instances of the official qa-context test set. Our system reaches a mean token accuracy of 0.9643 and produces accurate extractive spans. On the held-out split, fine-tuning roughly doubles both exact match (23.81 {\textrightarrow} 53.97) and token-level F1 (37.00 {\textrightarrow} 68.52) relative to the base model. The adapter corrects over-refusal tendency in addition to improving span precision. The base model emitted the prescribed no-answer string in 60{\%} of passages with answers, versus 0{\%} for the fine-tuned model. We report error analysis identifying recurring failures, and discuss the model{'}s weaker behaviour on cross-lingual context instances, which make up the large majority of the official test set, as a direct consequence of training exclusively on monolingual data."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="alkhder-etal-2026-monolingual">
<titleInfo>
<title>A Monolingual LoRA-Tuned Qwen2.5-3B System for Arabic Context-Based Question Answering</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hasan</namePart>
<namePart type="family">Alkhder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Hanini</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Imane</namePart>
<namePart type="family">Hocine</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maher</namePart>
<namePart type="family">Pasa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amro</namePart>
<namePart type="family">Najjar</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 Eleventh Conference on Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Koehn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christof</namePart>
<namePart type="family">Monz</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">979-8-89176-474-3</identifier>
</relatedItem>
<abstract>In this paper, we describe our submission to the WMT26 Multilingual Instruction Shared Task (MIST), Sub-task 1 (context-based question answering). We fine-tune Qwen2.5-3B-Instruct with a LoRA adapter on 3,112 extractive-QA examples drawn from the answerable subset of TyDi QA as released in the wmt26-mist-sample data. We submit outputs for the monolingual Arabic (question and context both in Arabic) instances of the official qa-context test set. Our system reaches a mean token accuracy of 0.9643 and produces accurate extractive spans. On the held-out split, fine-tuning roughly doubles both exact match (23.81 → 53.97) and token-level F1 (37.00 → 68.52) relative to the base model. The adapter corrects over-refusal tendency in addition to improving span precision. The base model emitted the prescribed no-answer string in 60% of passages with answers, versus 0% for the fine-tuned model. We report error analysis identifying recurring failures, and discuss the model’s weaker behaviour on cross-lingual context instances, which make up the large majority of the official test set, as a direct consequence of training exclusively on monolingual data.</abstract>
<identifier type="citekey">alkhder-etal-2026-monolingual</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.125/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1899</start>
<end>1905</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Monolingual LoRA-Tuned Qwen2.5-3B System for Arabic Context-Based Question Answering
%A Alkhder, Hasan
%A Hanini, Maria
%A Hocine, Imane
%A Pasa, Maher
%A Najjar, Amro
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Eleventh Conference on Machine Translation
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-474-3
%F alkhder-etal-2026-monolingual
%X In this paper, we describe our submission to the WMT26 Multilingual Instruction Shared Task (MIST), Sub-task 1 (context-based question answering). We fine-tune Qwen2.5-3B-Instruct with a LoRA adapter on 3,112 extractive-QA examples drawn from the answerable subset of TyDi QA as released in the wmt26-mist-sample data. We submit outputs for the monolingual Arabic (question and context both in Arabic) instances of the official qa-context test set. Our system reaches a mean token accuracy of 0.9643 and produces accurate extractive spans. On the held-out split, fine-tuning roughly doubles both exact match (23.81 → 53.97) and token-level F1 (37.00 → 68.52) relative to the base model. The adapter corrects over-refusal tendency in addition to improving span precision. The base model emitted the prescribed no-answer string in 60% of passages with answers, versus 0% for the fine-tuned model. We report error analysis identifying recurring failures, and discuss the model’s weaker behaviour on cross-lingual context instances, which make up the large majority of the official test set, as a direct consequence of training exclusively on monolingual data.
%U https://aclanthology.org/2026.wmt-main.125/
%P 1899-1905
Markdown (Informal)
[A Monolingual LoRA-Tuned Qwen2.5-3B System for Arabic Context-Based Question Answering](https://aclanthology.org/2026.wmt-main.125/) (Alkhder et al., WMT 2026)
ACL