@inproceedings{abdelaziz-etal-2026-casbai,
title = "{C}asb{AI} at {A}ra{S}ent{E}val 2026: Robust Dialectal {A}rabic Sentiment Classification via Multi-Seed Ensembling and Data Augmentation.",
author = "Abdelaziz, Chaima and
Saadaoui, KahinaHouda and
BELBACHIR, Faiza and
Said Lhadj, Lynda",
editor = "Al-Khalifa, Hend and
El-Haj, Mo and
Ezzini, Saad",
booktitle = "The 7th Workshop on Open-Source {A}rabic Corpora and Processing Tools ({OSACT}7) with 5 Shared Tasks",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.osact-1.38/",
doi = "10.63317/32cq3gywnhjs",
pages = "278--283",
abstract = "This paper describes the system we designed for our participation in the AraSentEval 2026 shared task on Arabic dialectal sentiment analysis. We propose a transformer-based approach relying on MARBERT combined with a multi-seed ensemble strategy and several optimization techniques. Our system integrates seven independently trained models with different random initializations and applies Stochastic Weight Averaging (SWA) to improve generalization. To address class imbalance, we augment the training data through dialectal synonym replacement, increasing the dataset size by 13.9{\%} while preserving dialect distribution. In addition, we incorporate Test-Time Augmentation (TTA) and investigate the use of pseudo-labeling based on high-confidence predictions. We report our experiments on the official dataset covering Moroccan, Egyptian, Jordanian, and Saudi dialects, and analyze the contribution of each component through ablation experiments. Our system achieved a macro F1-score of 84.62{\%} on the test set, ranking 3rd among 15 participating teams."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="abdelaziz-etal-2026-casbai">
<titleInfo>
<title>CasbAI at AraSentEval 2026: Robust Dialectal Arabic Sentiment Classification via Multi-Seed Ensembling and Data Augmentation.</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chaima</namePart>
<namePart type="family">Abdelaziz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">KahinaHouda</namePart>
<namePart type="family">Saadaoui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Faiza</namePart>
<namePart type="family">BELBACHIR</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lynda</namePart>
<namePart type="family">Said Lhadj</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>The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks</title>
</titleInfo>
<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">Mo</namePart>
<namePart type="family">El-Haj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saad</namePart>
<namePart type="family">Ezzini</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper describes the system we designed for our participation in the AraSentEval 2026 shared task on Arabic dialectal sentiment analysis. We propose a transformer-based approach relying on MARBERT combined with a multi-seed ensemble strategy and several optimization techniques. Our system integrates seven independently trained models with different random initializations and applies Stochastic Weight Averaging (SWA) to improve generalization. To address class imbalance, we augment the training data through dialectal synonym replacement, increasing the dataset size by 13.9% while preserving dialect distribution. In addition, we incorporate Test-Time Augmentation (TTA) and investigate the use of pseudo-labeling based on high-confidence predictions. We report our experiments on the official dataset covering Moroccan, Egyptian, Jordanian, and Saudi dialects, and analyze the contribution of each component through ablation experiments. Our system achieved a macro F1-score of 84.62% on the test set, ranking 3rd among 15 participating teams.</abstract>
<identifier type="citekey">abdelaziz-etal-2026-casbai</identifier>
<identifier type="doi">10.63317/32cq3gywnhjs</identifier>
<location>
<url>https://aclanthology.org/2026.osact-1.38/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>278</start>
<end>283</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CasbAI at AraSentEval 2026: Robust Dialectal Arabic Sentiment Classification via Multi-Seed Ensembling and Data Augmentation.
%A Abdelaziz, Chaima
%A Saadaoui, KahinaHouda
%A BELBACHIR, Faiza
%A Said Lhadj, Lynda
%Y Al-Khalifa, Hend
%Y El-Haj, Mo
%Y Ezzini, Saad
%S The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks
%D 2026
%8 May
%I Association for Computational Linguistics
%C Palma, Mallorca (Spain)
%F abdelaziz-etal-2026-casbai
%X This paper describes the system we designed for our participation in the AraSentEval 2026 shared task on Arabic dialectal sentiment analysis. We propose a transformer-based approach relying on MARBERT combined with a multi-seed ensemble strategy and several optimization techniques. Our system integrates seven independently trained models with different random initializations and applies Stochastic Weight Averaging (SWA) to improve generalization. To address class imbalance, we augment the training data through dialectal synonym replacement, increasing the dataset size by 13.9% while preserving dialect distribution. In addition, we incorporate Test-Time Augmentation (TTA) and investigate the use of pseudo-labeling based on high-confidence predictions. We report our experiments on the official dataset covering Moroccan, Egyptian, Jordanian, and Saudi dialects, and analyze the contribution of each component through ablation experiments. Our system achieved a macro F1-score of 84.62% on the test set, ranking 3rd among 15 participating teams.
%R 10.63317/32cq3gywnhjs
%U https://aclanthology.org/2026.osact-1.38/
%U https://doi.org/10.63317/32cq3gywnhjs
%P 278-283
Markdown (Informal)
[CasbAI at AraSentEval 2026: Robust Dialectal Arabic Sentiment Classification via Multi-Seed Ensembling and Data Augmentation.](https://aclanthology.org/2026.osact-1.38/) (Abdelaziz et al., OSACT 2026)
ACL