@inproceedings{ehab-etal-2026-raid,
title = "Raid {AI} at {A}r{A}-{DF} 2026: Data, Scale and Inference-Time Augmentation for {A}rabic Speech Deepfake Detection",
author = "Ehab, Ahmed and
Salem, Mooaz and
Mounir, Mohamed and
Abdelrazek, Seifeldin Hane{'}a Ismail",
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.195/",
pages = "1512--1518",
ISBN = "979-8-89176-497-2",
abstract = "We describe Raid AI{'}s entry to Track 1 (Dialect Generalization) of the ArA-DF 2026 shared task on Arabic speech deepfake detection, which placed 1st of 15 teams at 0.8719{\%} Equal Error Rate against a 14.54{\%} official baseline. The configuration of the submitted system is proprietary and is not disclosed. We instead report the ablation behind it, re-scored on the official 144,210-clip evaluation partition against the labels released after the challenge, with bootstrap confidence intervals and paired tests throughout. On development-test, corpus and augmentation alone move a fixed architecture from 13.13{\%} to 1.37{\%} EER. Encoder scale is not monotonic: a 1B encoder beats a 2B under a recipe matched on corpus, augmentation, learning rate, class weights and effective batch size, though the gap is inside the run-to-run spread we measure from three seeds. Three-crop test-time augmentation improves all ten models tested, by 17.4{\%} relative on average, and is the only effect immune to seed variance because it is measured within a checkpoint; a per-crop decomposition shows the gain comes from averaging rather than from choosing a better window."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ehab-etal-2026-raid">
<titleInfo>
<title>Raid AI at ArA-DF 2026: Data, Scale and Inference-Time Augmentation for Arabic Speech Deepfake Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Ehab</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mooaz</namePart>
<namePart type="family">Salem</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohamed</namePart>
<namePart type="family">Mounir</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seifeldin</namePart>
<namePart type="given">Hane’a</namePart>
<namePart type="given">Ismail</namePart>
<namePart type="family">Abdelrazek</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>We describe Raid AI’s entry to Track 1 (Dialect Generalization) of the ArA-DF 2026 shared task on Arabic speech deepfake detection, which placed 1st of 15 teams at 0.8719% Equal Error Rate against a 14.54% official baseline. The configuration of the submitted system is proprietary and is not disclosed. We instead report the ablation behind it, re-scored on the official 144,210-clip evaluation partition against the labels released after the challenge, with bootstrap confidence intervals and paired tests throughout. On development-test, corpus and augmentation alone move a fixed architecture from 13.13% to 1.37% EER. Encoder scale is not monotonic: a 1B encoder beats a 2B under a recipe matched on corpus, augmentation, learning rate, class weights and effective batch size, though the gap is inside the run-to-run spread we measure from three seeds. Three-crop test-time augmentation improves all ten models tested, by 17.4% relative on average, and is the only effect immune to seed variance because it is measured within a checkpoint; a per-crop decomposition shows the gain comes from averaging rather than from choosing a better window.</abstract>
<identifier type="citekey">ehab-etal-2026-raid</identifier>
<location>
<url>https://aclanthology.org/2026.arabicnlp-sharedtasks.195/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1512</start>
<end>1518</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Raid AI at ArA-DF 2026: Data, Scale and Inference-Time Augmentation for Arabic Speech Deepfake Detection
%A Ehab, Ahmed
%A Salem, Mooaz
%A Mounir, Mohamed
%A Abdelrazek, Seifeldin Hane’a Ismail
%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 ehab-etal-2026-raid
%X We describe Raid AI’s entry to Track 1 (Dialect Generalization) of the ArA-DF 2026 shared task on Arabic speech deepfake detection, which placed 1st of 15 teams at 0.8719% Equal Error Rate against a 14.54% official baseline. The configuration of the submitted system is proprietary and is not disclosed. We instead report the ablation behind it, re-scored on the official 144,210-clip evaluation partition against the labels released after the challenge, with bootstrap confidence intervals and paired tests throughout. On development-test, corpus and augmentation alone move a fixed architecture from 13.13% to 1.37% EER. Encoder scale is not monotonic: a 1B encoder beats a 2B under a recipe matched on corpus, augmentation, learning rate, class weights and effective batch size, though the gap is inside the run-to-run spread we measure from three seeds. Three-crop test-time augmentation improves all ten models tested, by 17.4% relative on average, and is the only effect immune to seed variance because it is measured within a checkpoint; a per-crop decomposition shows the gain comes from averaging rather than from choosing a better window.
%U https://aclanthology.org/2026.arabicnlp-sharedtasks.195/
%P 1512-1518
Markdown (Informal)
[Raid AI at ArA-DF 2026: Data, Scale and Inference-Time Augmentation for Arabic Speech Deepfake Detection](https://aclanthology.org/2026.arabicnlp-sharedtasks.195/) (Ehab et al., ArabicNLP 2026)
ACL