<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="udagawa-etal-2025-bias">
    <titleInfo>
        <title>Bias Analysis and Mitigation through Protected Attribute Detection and Regard Classification</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Takuma</namePart>
        <namePart type="family">Udagawa</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yang</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hiroshi</namePart>
        <namePart type="family">Kanayama</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Bishwaranjan</namePart>
        <namePart type="family">Bhattacharjee</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Christos</namePart>
            <namePart type="family">Christodoulopoulos</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tanmoy</namePart>
            <namePart type="family">Chakraborty</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Carolyn</namePart>
            <namePart type="family">Rose</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Violet</namePart>
            <namePart type="family">Peng</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Suzhou, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-335-7</identifier>
    </relatedItem>
    <abstract>Large language models (LLMs) acquire general linguistic knowledge from massive-scale pretraining. However, pretraining data mainly comprised of web-crawled texts contain undesirable social biases which can be perpetuated or even amplified by LLMs. In this study, we propose an efficient yet effective annotation pipeline to investigate social biases in the pretraining corpora. Our pipeline consists of protected attribute detection to identify diverse demographics, followed by regard classification to analyze the language polarity towards each attribute. Through our experiments, we demonstrate the effect of our bias analysis and mitigation measures, focusing on Common Crawl as the most representative pretraining corpus.</abstract>
    <identifier type="citekey">udagawa-etal-2025-bias</identifier>
    <location>
        <url>https://aclanthology.org/2025.findings-emnlp.2/</url>
    </location>
    <part>
        <date>2025-11</date>
        <extent unit="page">
            <start>16</start>
            <end>25</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Bias Analysis and Mitigation through Protected Attribute Detection and Regard Classification
%A Udagawa, Takuma
%A Zhao, Yang
%A Kanayama, Hiroshi
%A Bhattacharjee, Bishwaranjan
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F udagawa-etal-2025-bias
%X Large language models (LLMs) acquire general linguistic knowledge from massive-scale pretraining. However, pretraining data mainly comprised of web-crawled texts contain undesirable social biases which can be perpetuated or even amplified by LLMs. In this study, we propose an efficient yet effective annotation pipeline to investigate social biases in the pretraining corpora. Our pipeline consists of protected attribute detection to identify diverse demographics, followed by regard classification to analyze the language polarity towards each attribute. Through our experiments, we demonstrate the effect of our bias analysis and mitigation measures, focusing on Common Crawl as the most representative pretraining corpus.
%U https://aclanthology.org/2025.findings-emnlp.2/
%P 16-25

Markdown (Informal)

[Bias Analysis and Mitigation through Protected Attribute Detection and Regard Classification](https://aclanthology.org/2025.findings-emnlp.2/) (Udagawa et al., Findings 2025)

• Bias Analysis and Mitigation through Protected Attribute Detection and Regard Classification (Udagawa et al., Findings 2025)

ACL

• Takuma Udagawa, Yang Zhao, Hiroshi Kanayama, and Bishwaranjan Bhattacharjee. 2025. Bias Analysis and Mitigation through Protected Attribute Detection and Regard Classification. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 16–25, Suzhou, China. Association for Computational Linguistics.