@inproceedings{kim-etal-2025-analyzing,
title = "Analyzing Offensive Language Dataset Insights from Training Dynamics and Human Agreement Level",
author = "Kim, Do Kyung and
Ahn, Hyeseon and
Kim, Youngwook and
Han, Yo-Sub",
editor = "Rambow, Owen and
Wanner, Leo and
Apidianaki, Marianna and
Al-Khalifa, Hend and
Eugenio, Barbara Di and
Schockaert, Steven",
booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
month = jan,
year = "2025",
address = "Abu Dhabi, UAE",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.coling-main.653/",
pages = "9780--9792",
abstract = "Implicit hate speech detection is challenging due to its subjectivity and context dependence, with existing models often struggling in outof-domain scenarios. We propose CONELA, a novel data refinement strategy that enhances model performance and generalization by integrating human annotation agreement with model training dynamics. By removing both easy and hard instances from the model`s perspective, while also considering whether humans agree or disagree and retaining ambiguous cases crucial for out-of-distribution generalization, CONELA consistently improves performance across multiple datasets and models. We also observe significant improvements in F1 scores and cross-domain generalization with the use of our CONELA strategy. Addressing data scarcity in smaller datasets, we introduce a weighted loss function and an ensemble strategy incorporating disagreement maximization, effectively balancing learning from limited data. Our findings demonstrate that refining datasets by integrating both model and human perspectives significantly enhances the effectiveness and generalization of implicit hate speech detection models. This approach lays a strong foundation for future research on dataset refinement and model robustness."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kim-etal-2025-analyzing">
<titleInfo>
<title>Analyzing Offensive Language Dataset Insights from Training Dynamics and Human Agreement Level</title>
</titleInfo>
<name type="personal">
<namePart type="given">Do</namePart>
<namePart type="given">Kyung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hyeseon</namePart>
<namePart type="family">Ahn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Youngwook</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yo-Sub</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-01</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 31st International Conference on Computational Linguistics</title>
</titleInfo>
<name type="personal">
<namePart type="given">Owen</namePart>
<namePart type="family">Rambow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leo</namePart>
<namePart type="family">Wanner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marianna</namePart>
<namePart type="family">Apidianaki</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">Barbara</namePart>
<namePart type="given">Di</namePart>
<namePart type="family">Eugenio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Schockaert</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, UAE</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Implicit hate speech detection is challenging due to its subjectivity and context dependence, with existing models often struggling in outof-domain scenarios. We propose CONELA, a novel data refinement strategy that enhances model performance and generalization by integrating human annotation agreement with model training dynamics. By removing both easy and hard instances from the model‘s perspective, while also considering whether humans agree or disagree and retaining ambiguous cases crucial for out-of-distribution generalization, CONELA consistently improves performance across multiple datasets and models. We also observe significant improvements in F1 scores and cross-domain generalization with the use of our CONELA strategy. Addressing data scarcity in smaller datasets, we introduce a weighted loss function and an ensemble strategy incorporating disagreement maximization, effectively balancing learning from limited data. Our findings demonstrate that refining datasets by integrating both model and human perspectives significantly enhances the effectiveness and generalization of implicit hate speech detection models. This approach lays a strong foundation for future research on dataset refinement and model robustness.</abstract>
<identifier type="citekey">kim-etal-2025-analyzing</identifier>
<location>
<url>https://aclanthology.org/2025.coling-main.653/</url>
</location>
<part>
<date>2025-01</date>
<extent unit="page">
<start>9780</start>
<end>9792</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Analyzing Offensive Language Dataset Insights from Training Dynamics and Human Agreement Level
%A Kim, Do Kyung
%A Ahn, Hyeseon
%A Kim, Youngwook
%A Han, Yo-Sub
%Y Rambow, Owen
%Y Wanner, Leo
%Y Apidianaki, Marianna
%Y Al-Khalifa, Hend
%Y Eugenio, Barbara Di
%Y Schockaert, Steven
%S Proceedings of the 31st International Conference on Computational Linguistics
%D 2025
%8 January
%I Association for Computational Linguistics
%C Abu Dhabi, UAE
%F kim-etal-2025-analyzing
%X Implicit hate speech detection is challenging due to its subjectivity and context dependence, with existing models often struggling in outof-domain scenarios. We propose CONELA, a novel data refinement strategy that enhances model performance and generalization by integrating human annotation agreement with model training dynamics. By removing both easy and hard instances from the model‘s perspective, while also considering whether humans agree or disagree and retaining ambiguous cases crucial for out-of-distribution generalization, CONELA consistently improves performance across multiple datasets and models. We also observe significant improvements in F1 scores and cross-domain generalization with the use of our CONELA strategy. Addressing data scarcity in smaller datasets, we introduce a weighted loss function and an ensemble strategy incorporating disagreement maximization, effectively balancing learning from limited data. Our findings demonstrate that refining datasets by integrating both model and human perspectives significantly enhances the effectiveness and generalization of implicit hate speech detection models. This approach lays a strong foundation for future research on dataset refinement and model robustness.
%U https://aclanthology.org/2025.coling-main.653/
%P 9780-9792
Markdown (Informal)
[Analyzing Offensive Language Dataset Insights from Training Dynamics and Human Agreement Level](https://aclanthology.org/2025.coling-main.653/) (Kim et al., COLING 2025)
ACL