@inproceedings{nhan-thin-2026-gradient,
title = "Gradient Descenders at {S}em{E}val-2026 Task 9: Data-Centric Counterfactual Augmentation for Multi-Label Hate Speech Detection",
author = "Nhan, Tran and
Thin, Dang",
editor = "Kochmar, Ekaterina and
Ghosh, Debanjan and
North, Kai and
Komachi, Mamoru",
booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.semeval-1.113/",
pages = "811--820",
ISBN = "979-8-89176-414-9",
abstract = "In this paper, we describe the Gradient Descenders submission to SemEval-2026 Task 9 Subtask 2: Multi-Label Hate Speech Detection. Existing Transformer-based approaches often exhibit degraded performance on this task due to severe class imbalance and complex class intersectionality, leading to the learning of spurious correlations. To counteract this, we introduce a novel, data-centric counterfactual augmentation pipeline. We employ Large Language Models (LLMs) as semantic generators to synthesize diverse, targeted training samples via three distinct prompting strategies: Additive Label-Flipping (Attribute Injection), Context Decoupling, and Cross-Domain Identity Substitution. Fine-tuning a RoBERTa classifier on this augmented corpus significantly improves the model{'}s sensitivity to minority classes. Ultimately, our system achieves a Macro-F1 score of 44.15{\textbackslash}{\%} on the official test set, highlighting the efficacy of targeted LLM-based augmentation in highly imbalanced, multi-label environments."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nhan-thin-2026-gradient">
<titleInfo>
<title>Gradient Descenders at SemEval-2026 Task 9: Data-Centric Counterfactual Augmentation for Multi-Label Hate Speech Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Tran</namePart>
<namePart type="family">Nhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dang</namePart>
<namePart type="family">Thin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 20th International Workshop on Semantic Evaluation (2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Kochmar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Debanjan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai</namePart>
<namePart type="family">North</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mamoru</namePart>
<namePart type="family">Komachi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-414-9</identifier>
</relatedItem>
<abstract>In this paper, we describe the Gradient Descenders submission to SemEval-2026 Task 9 Subtask 2: Multi-Label Hate Speech Detection. Existing Transformer-based approaches often exhibit degraded performance on this task due to severe class imbalance and complex class intersectionality, leading to the learning of spurious correlations. To counteract this, we introduce a novel, data-centric counterfactual augmentation pipeline. We employ Large Language Models (LLMs) as semantic generators to synthesize diverse, targeted training samples via three distinct prompting strategies: Additive Label-Flipping (Attribute Injection), Context Decoupling, and Cross-Domain Identity Substitution. Fine-tuning a RoBERTa classifier on this augmented corpus significantly improves the model’s sensitivity to minority classes. Ultimately, our system achieves a Macro-F1 score of 44.15\textbackslash% on the official test set, highlighting the efficacy of targeted LLM-based augmentation in highly imbalanced, multi-label environments.</abstract>
<identifier type="citekey">nhan-thin-2026-gradient</identifier>
<location>
<url>https://aclanthology.org/2026.semeval-1.113/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>811</start>
<end>820</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Gradient Descenders at SemEval-2026 Task 9: Data-Centric Counterfactual Augmentation for Multi-Label Hate Speech Detection
%A Nhan, Tran
%A Thin, Dang
%Y Kochmar, Ekaterina
%Y Ghosh, Debanjan
%Y North, Kai
%Y Komachi, Mamoru
%S Proceedings of the 20th International Workshop on Semantic Evaluation (2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-414-9
%F nhan-thin-2026-gradient
%X In this paper, we describe the Gradient Descenders submission to SemEval-2026 Task 9 Subtask 2: Multi-Label Hate Speech Detection. Existing Transformer-based approaches often exhibit degraded performance on this task due to severe class imbalance and complex class intersectionality, leading to the learning of spurious correlations. To counteract this, we introduce a novel, data-centric counterfactual augmentation pipeline. We employ Large Language Models (LLMs) as semantic generators to synthesize diverse, targeted training samples via three distinct prompting strategies: Additive Label-Flipping (Attribute Injection), Context Decoupling, and Cross-Domain Identity Substitution. Fine-tuning a RoBERTa classifier on this augmented corpus significantly improves the model’s sensitivity to minority classes. Ultimately, our system achieves a Macro-F1 score of 44.15\textbackslash% on the official test set, highlighting the efficacy of targeted LLM-based augmentation in highly imbalanced, multi-label environments.
%U https://aclanthology.org/2026.semeval-1.113/
%P 811-820
Markdown (Informal)
[Gradient Descenders at SemEval-2026 Task 9: Data-Centric Counterfactual Augmentation for Multi-Label Hate Speech Detection](https://aclanthology.org/2026.semeval-1.113/) (Nhan & Thin, SemEval 2026)
ACL