@inproceedings{basso-madjoukeng-etal-2026-leveraging,
title = "Leveraging Unannotated Sign Language Data via a Robust Data Augmentation Method for Contrastive Representation Learning",
author = "Basso Madjoukeng, Ariel and
Poitier, Pierre and
Kenmogne, Belise Edith and
Couplet, Adelaide and
Leleu, Margaux and
Benoit, Frenay",
editor = "Efthimiou, Eleni and
Fotinea, Stavroula-Evita and
Hanke, Thomas and
Hochgesang, Julie A. and
Mesch, Johanna and
Schulder, Marc",
booktitle = "Proceedings of the {LREC} 2026 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.signlang-1.2/",
doi = "10.63317/4qnrmhtbv9cz",
pages = "10--16",
abstract = "Contrastive learning is a deep learning paradigm that allows the learning of useful representations without annotations. In many fields, including sign language recognition (SLR), contrastive approaches have proven to be very effective for developing pretrained models. To learn representations, they generate augmented variants of an instance through augmentation techniques and then maximize their similarities. The quality of the learned representations is strongly correlated with the augmentations used during training. In several fields, specialized augmentations have been developed and adopted. However, in SLR, we observed two trends: contrastive-based SLR approaches often rely on augmentations that are not realistic for the application (e.g., vertical flip, excessive rotations); specialized augmentation methods lack robustness. Hence, when they are used as a starting point for contrastive algorithms, the learned representations are often irrelevant, and sometimes sensitive. These issues considerably affect the accuracy of SLR models on downstream tasks. In response, this paper proposes a robust augmentation method specially designed for contrastive approaches applied to SLR. The results show an improvement in accuracy during linear evaluation and semi-supervised learning with only 30{\%} of annotations."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="basso-madjoukeng-etal-2026-leveraging">
<titleInfo>
<title>Leveraging Unannotated Sign Language Data via a Robust Data Augmentation Method for Contrastive Representation Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ariel</namePart>
<namePart type="family">Basso Madjoukeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pierre</namePart>
<namePart type="family">Poitier</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Belise</namePart>
<namePart type="given">Edith</namePart>
<namePart type="family">Kenmogne</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adelaide</namePart>
<namePart type="family">Couplet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Margaux</namePart>
<namePart type="family">Leleu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Frenay</namePart>
<namePart type="family">Benoit</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>Proceedings of the LREC 2026 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion</title>
</titleInfo>
<name type="personal">
<namePart type="given">Eleni</namePart>
<namePart type="family">Efthimiou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stavroula-Evita</namePart>
<namePart type="family">Fotinea</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Hanke</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Julie</namePart>
<namePart type="given">A</namePart>
<namePart type="family">Hochgesang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Johanna</namePart>
<namePart type="family">Mesch</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marc</namePart>
<namePart type="family">Schulder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Contrastive learning is a deep learning paradigm that allows the learning of useful representations without annotations. In many fields, including sign language recognition (SLR), contrastive approaches have proven to be very effective for developing pretrained models. To learn representations, they generate augmented variants of an instance through augmentation techniques and then maximize their similarities. The quality of the learned representations is strongly correlated with the augmentations used during training. In several fields, specialized augmentations have been developed and adopted. However, in SLR, we observed two trends: contrastive-based SLR approaches often rely on augmentations that are not realistic for the application (e.g., vertical flip, excessive rotations); specialized augmentation methods lack robustness. Hence, when they are used as a starting point for contrastive algorithms, the learned representations are often irrelevant, and sometimes sensitive. These issues considerably affect the accuracy of SLR models on downstream tasks. In response, this paper proposes a robust augmentation method specially designed for contrastive approaches applied to SLR. The results show an improvement in accuracy during linear evaluation and semi-supervised learning with only 30% of annotations.</abstract>
<identifier type="citekey">basso-madjoukeng-etal-2026-leveraging</identifier>
<identifier type="doi">10.63317/4qnrmhtbv9cz</identifier>
<location>
<url>https://aclanthology.org/2026.signlang-1.2/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>10</start>
<end>16</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Leveraging Unannotated Sign Language Data via a Robust Data Augmentation Method for Contrastive Representation Learning
%A Basso Madjoukeng, Ariel
%A Poitier, Pierre
%A Kenmogne, Belise Edith
%A Couplet, Adelaide
%A Leleu, Margaux
%A Benoit, Frenay
%Y Efthimiou, Eleni
%Y Fotinea, Stavroula-Evita
%Y Hanke, Thomas
%Y Hochgesang, Julie A.
%Y Mesch, Johanna
%Y Schulder, Marc
%S Proceedings of the LREC 2026 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F basso-madjoukeng-etal-2026-leveraging
%X Contrastive learning is a deep learning paradigm that allows the learning of useful representations without annotations. In many fields, including sign language recognition (SLR), contrastive approaches have proven to be very effective for developing pretrained models. To learn representations, they generate augmented variants of an instance through augmentation techniques and then maximize their similarities. The quality of the learned representations is strongly correlated with the augmentations used during training. In several fields, specialized augmentations have been developed and adopted. However, in SLR, we observed two trends: contrastive-based SLR approaches often rely on augmentations that are not realistic for the application (e.g., vertical flip, excessive rotations); specialized augmentation methods lack robustness. Hence, when they are used as a starting point for contrastive algorithms, the learned representations are often irrelevant, and sometimes sensitive. These issues considerably affect the accuracy of SLR models on downstream tasks. In response, this paper proposes a robust augmentation method specially designed for contrastive approaches applied to SLR. The results show an improvement in accuracy during linear evaluation and semi-supervised learning with only 30% of annotations.
%R 10.63317/4qnrmhtbv9cz
%U https://aclanthology.org/2026.signlang-1.2/
%U https://doi.org/10.63317/4qnrmhtbv9cz
%P 10-16
Markdown (Informal)
[Leveraging Unannotated Sign Language Data via a Robust Data Augmentation Method for Contrastive Representation Learning](https://aclanthology.org/2026.signlang-1.2/) (Basso Madjoukeng et al., SignLang 2026)
ACL