@inproceedings{s-etal-2025-wise,
title = "Wise@{LT}-{EDI}-2025: Combining Classical and Neural Representations with Multi-scale Ensemble Learning for Code-mixed Hate Speech Detection",
author = "S, Ganesh Sundhar and
K, Durai Singh and
G, Gnanasabesan and
N, Hari Krishnan and
Dhanush, Mc",
editor = "Gkirtzou, Katerina and
{\v{Z}}itnik, Slavko and
Gracia, Jorge and
Gromann, Dagmar and
di Buono, Maria Pia and
Monti, Johanna and
Ionov, Maxim",
booktitle = "Proceedings of the 5th Conference on Language, Data and Knowledge: Fifth Workshop on Language Technology for Equality, Diversity, Inclusion",
month = sep,
year = "2025",
address = "Naples, Italy",
publisher = "Unior Press",
url = "https://aclanthology.org/2025.ltedi-1.9/",
pages = "54--62",
ISBN = "978-88-6719-334-9",
abstract = "Detecting hate speech targeting caste and migration communities in code-mixed Tamil-English social media content is challenging due to limited resources and socio-cultural complexities. This paper proposes a multi-scale hybrid architecture combining classical and neural representations with hierarchical ensemble learning. We employ advanced preprocessing including transliteration and character repetition removal, then extract features using classical TF-IDF vectors at multiple scales (512, 1024, 2048) processed through linear layers, alongside contextual embeddings from five transformer models-Google BERT, XLM-RoBERTa (Base and Large), SeanBenhur BERT, and IndicBERT. These concatenated representations encode both statistical and contextual information, which are input to multiple ML classification heads (Random Forest, SVM, etc). A three-level hierarchical ensemble strategy combines predictions across classifiers, transformer-TF-IDF combinations, and dimensional scales for enhanced robustness. Our method scored an F1-score of 0.818, ranking 3rd in the LT-EDI-2025 shared task, showing the efficacy of blending classical and neural methods with multi-level ensemble learning for hate speech detection in low-resource languages."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="s-etal-2025-wise">
<titleInfo>
<title>Wise@LT-EDI-2025: Combining Classical and Neural Representations with Multi-scale Ensemble Learning for Code-mixed Hate Speech Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ganesh</namePart>
<namePart type="given">Sundhar</namePart>
<namePart type="family">S</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Durai</namePart>
<namePart type="given">Singh</namePart>
<namePart type="family">K</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gnanasabesan</namePart>
<namePart type="family">G</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hari</namePart>
<namePart type="given">Krishnan</namePart>
<namePart type="family">N</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mc</namePart>
<namePart type="family">Dhanush</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-09</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 5th Conference on Language, Data and Knowledge: Fifth Workshop on Language Technology for Equality, Diversity, Inclusion</title>
</titleInfo>
<name type="personal">
<namePart type="given">Katerina</namePart>
<namePart type="family">Gkirtzou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Slavko</namePart>
<namePart type="family">Žitnik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jorge</namePart>
<namePart type="family">Gracia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dagmar</namePart>
<namePart type="family">Gromann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="given">Pia</namePart>
<namePart type="family">di Buono</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Johanna</namePart>
<namePart type="family">Monti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maxim</namePart>
<namePart type="family">Ionov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Unior Press</publisher>
<place>
<placeTerm type="text">Naples, Italy</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">978-88-6719-334-9</identifier>
</relatedItem>
<abstract>Detecting hate speech targeting caste and migration communities in code-mixed Tamil-English social media content is challenging due to limited resources and socio-cultural complexities. This paper proposes a multi-scale hybrid architecture combining classical and neural representations with hierarchical ensemble learning. We employ advanced preprocessing including transliteration and character repetition removal, then extract features using classical TF-IDF vectors at multiple scales (512, 1024, 2048) processed through linear layers, alongside contextual embeddings from five transformer models-Google BERT, XLM-RoBERTa (Base and Large), SeanBenhur BERT, and IndicBERT. These concatenated representations encode both statistical and contextual information, which are input to multiple ML classification heads (Random Forest, SVM, etc). A three-level hierarchical ensemble strategy combines predictions across classifiers, transformer-TF-IDF combinations, and dimensional scales for enhanced robustness. Our method scored an F1-score of 0.818, ranking 3rd in the LT-EDI-2025 shared task, showing the efficacy of blending classical and neural methods with multi-level ensemble learning for hate speech detection in low-resource languages.</abstract>
<identifier type="citekey">s-etal-2025-wise</identifier>
<location>
<url>https://aclanthology.org/2025.ltedi-1.9/</url>
</location>
<part>
<date>2025-09</date>
<extent unit="page">
<start>54</start>
<end>62</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Wise@LT-EDI-2025: Combining Classical and Neural Representations with Multi-scale Ensemble Learning for Code-mixed Hate Speech Detection
%A S, Ganesh Sundhar
%A K, Durai Singh
%A G, Gnanasabesan
%A N, Hari Krishnan
%A Dhanush, Mc
%Y Gkirtzou, Katerina
%Y Žitnik, Slavko
%Y Gracia, Jorge
%Y Gromann, Dagmar
%Y di Buono, Maria Pia
%Y Monti, Johanna
%Y Ionov, Maxim
%S Proceedings of the 5th Conference on Language, Data and Knowledge: Fifth Workshop on Language Technology for Equality, Diversity, Inclusion
%D 2025
%8 September
%I Unior Press
%C Naples, Italy
%@ 978-88-6719-334-9
%F s-etal-2025-wise
%X Detecting hate speech targeting caste and migration communities in code-mixed Tamil-English social media content is challenging due to limited resources and socio-cultural complexities. This paper proposes a multi-scale hybrid architecture combining classical and neural representations with hierarchical ensemble learning. We employ advanced preprocessing including transliteration and character repetition removal, then extract features using classical TF-IDF vectors at multiple scales (512, 1024, 2048) processed through linear layers, alongside contextual embeddings from five transformer models-Google BERT, XLM-RoBERTa (Base and Large), SeanBenhur BERT, and IndicBERT. These concatenated representations encode both statistical and contextual information, which are input to multiple ML classification heads (Random Forest, SVM, etc). A three-level hierarchical ensemble strategy combines predictions across classifiers, transformer-TF-IDF combinations, and dimensional scales for enhanced robustness. Our method scored an F1-score of 0.818, ranking 3rd in the LT-EDI-2025 shared task, showing the efficacy of blending classical and neural methods with multi-level ensemble learning for hate speech detection in low-resource languages.
%U https://aclanthology.org/2025.ltedi-1.9/
%P 54-62
Markdown (Informal)
[Wise@LT-EDI-2025: Combining Classical and Neural Representations with Multi-scale Ensemble Learning for Code-mixed Hate Speech Detection](https://aclanthology.org/2025.ltedi-1.9/) (S et al., LTEDI 2025)
ACL
- Ganesh Sundhar S, Durai Singh K, Gnanasabesan G, Hari Krishnan N, and Mc Dhanush. 2025. Wise@LT-EDI-2025: Combining Classical and Neural Representations with Multi-scale Ensemble Learning for Code-mixed Hate Speech Detection. In Proceedings of the 5th Conference on Language, Data and Knowledge: Fifth Workshop on Language Technology for Equality, Diversity, Inclusion, pages 54–62, Naples, Italy. Unior Press.