@inproceedings{ibrahim-etal-2026-alexis-dialectsenteval,
title = "{ALEXIS} at {D}ialect{S}ent{E}val Subtask 1 Adversarial Training and Iterative Self-Training for Zero-Shot Dialect Sentiment Analysis",
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.187/",
pages = "1462--1466",
ISBN = "979-8-89176-497-2",
abstract = "Dialect sentiment analysis asks for the polarity of a sentence written in a regional Arabic variety rather than Modern Standard Arabic. We present a system for Subtask 1 of DialectSentEval built on two observations about the data. First, the task is one of zero-shot transfer rather than classification: the test set introduces Lebanese, a dialect with no labeled training data and the highest out-of-vocabulary rate in the task (52.0{\%} of word types, against 19.5{--}41.0{\%} for the seen dialects). Second, $k$-fold cross-validation cannot see this, since it measures generalization to unseen sentences and never to an unseen dialect. We therefore validate with a leave-one-dialect-out diagnostic and score every component on the official leaderboard. Three dialectally diverse encoders (MARBERTv2, CAMeLBERT-DA, AraBERTv2) are trained with Fast Gradient Method adversarial training, then refined by confidence-filtered self-training on the unlabeled test set and character-level noise augmentation. Adversarial training contributes $+0.040$ Macro F1 over a gold-only ensemble, self-training a further $+0.027$, and per-class pseudo-label filtering $+0.018$. The final system reaches 0.8667 Macro F1, ranking 1st of 12 submissions, 4.6 points above the runner-up."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ibrahim-etal-2026-alexis-dialectsenteval">
<titleInfo>
<title>ALEXIS at DialectSentEval Subtask 1 Adversarial Training and Iterative Self-Training for Zero-Shot Dialect Sentiment Analysis</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>Dialect sentiment analysis asks for the polarity of a sentence written in a regional Arabic variety rather than Modern Standard Arabic. We present a system for Subtask 1 of DialectSentEval built on two observations about the data. First, the task is one of zero-shot transfer rather than classification: the test set introduces Lebanese, a dialect with no labeled training data and the highest out-of-vocabulary rate in the task (52.0% of word types, against 19.5–41.0% for the seen dialects). Second, k-fold cross-validation cannot see this, since it measures generalization to unseen sentences and never to an unseen dialect. We therefore validate with a leave-one-dialect-out diagnostic and score every component on the official leaderboard. Three dialectally diverse encoders (MARBERTv2, CAMeLBERT-DA, AraBERTv2) are trained with Fast Gradient Method adversarial training, then refined by confidence-filtered self-training on the unlabeled test set and character-level noise augmentation. Adversarial training contributes +0.040 Macro F1 over a gold-only ensemble, self-training a further +0.027, and per-class pseudo-label filtering +0.018. The final system reaches 0.8667 Macro F1, ranking 1st of 12 submissions, 4.6 points above the runner-up.</abstract>
<identifier type="citekey">ibrahim-etal-2026-alexis-dialectsenteval</identifier>
<location>
<url>https://aclanthology.org/2026.arabicnlp-sharedtasks.187/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1462</start>
<end>1466</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings %T ALEXIS at DialectSentEval Subtask 1 Adversarial Training and Iterative Self-Training for Zero-Shot Dialect Sentiment Analysis %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-dialectsenteval %X Dialect sentiment analysis asks for the polarity of a sentence written in a regional Arabic variety rather than Modern Standard Arabic. We present a system for Subtask 1 of DialectSentEval built on two observations about the data. First, the task is one of zero-shot transfer rather than classification: the test set introduces Lebanese, a dialect with no labeled training data and the highest out-of-vocabulary rate in the task (52.0% of word types, against 19.5–41.0% for the seen dialects). Second, k-fold cross-validation cannot see this, since it measures generalization to unseen sentences and never to an unseen dialect. We therefore validate with a leave-one-dialect-out diagnostic and score every component on the official leaderboard. Three dialectally diverse encoders (MARBERTv2, CAMeLBERT-DA, AraBERTv2) are trained with Fast Gradient Method adversarial training, then refined by confidence-filtered self-training on the unlabeled test set and character-level noise augmentation. Adversarial training contributes +0.040 Macro F1 over a gold-only ensemble, self-training a further +0.027, and per-class pseudo-label filtering +0.018. The final system reaches 0.8667 Macro F1, ranking 1st of 12 submissions, 4.6 points above the runner-up. %U https://aclanthology.org/2026.arabicnlp-sharedtasks.187/ %P 1462-1466
Markdown (Informal)
[ALEXIS at DialectSentEval Subtask 1 Adversarial Training and Iterative Self-Training for Zero-Shot Dialect Sentiment Analysis](https://aclanthology.org/2026.arabicnlp-sharedtasks.187/) (Ibrahim et al., ArabicNLP 2026)