<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2025-preserving">
    <titleInfo>
        <title>Preserving Zero-shot Capability in Supervised Fine-tuning for Multi-label Text Classification</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Si-An</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hsuan-Tien</namePart>
        <namePart type="family">Lin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Chih-Jen</namePart>
        <namePart type="family">Lin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-04</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: NAACL 2025</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Luis</namePart>
            <namePart type="family">Chiruzzo</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alan</namePart>
            <namePart type="family">Ritter</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Lu</namePart>
            <namePart type="family">Wang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Albuquerque, New Mexico</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-195-7</identifier>
    </relatedItem>
    <abstract>Zero-shot multi-label text classification (ZMTC) requires models to predict multiple labels for a document, including labels unseen during training. Previous work assumes that models leveraging label descriptions ensures zero-shot capability. However, we find that supervised methods, despite achieving strong overall performance, lose their zero-shot capability during training, revealing a trade-off between overall and zero-shot performance. To address the issue, we propose OF-DE and OF-LAN, which preserve the zero-shot capabilities of powerful dual encoder and label-wise attention network architectures by freezing the label encoder. Additionally, we introduce a self-supervised auxiliary loss to further improve zero-shot performance. Experiments demonstrate that our approach significantly improves zero-shot performance of supervised methods while maintaining strong overall accuracy.</abstract>
    <identifier type="citekey">chen-etal-2025-preserving</identifier>
    <identifier type="doi">10.18653/v1/2025.findings-naacl.315</identifier>
    <location>
        <url>https://aclanthology.org/2025.findings-naacl.315/</url>
    </location>
    <part>
        <date>2025-04</date>
        <extent unit="page">
            <start>5699</start>
            <end>5712</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Preserving Zero-shot Capability in Supervised Fine-tuning for Multi-label Text Classification
%A Chen, Si-An
%A Lin, Hsuan-Tien
%A Lin, Chih-Jen
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Findings of the Association for Computational Linguistics: NAACL 2025
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-195-7
%F chen-etal-2025-preserving
%X Zero-shot multi-label text classification (ZMTC) requires models to predict multiple labels for a document, including labels unseen during training. Previous work assumes that models leveraging label descriptions ensures zero-shot capability. However, we find that supervised methods, despite achieving strong overall performance, lose their zero-shot capability during training, revealing a trade-off between overall and zero-shot performance. To address the issue, we propose OF-DE and OF-LAN, which preserve the zero-shot capabilities of powerful dual encoder and label-wise attention network architectures by freezing the label encoder. Additionally, we introduce a self-supervised auxiliary loss to further improve zero-shot performance. Experiments demonstrate that our approach significantly improves zero-shot performance of supervised methods while maintaining strong overall accuracy.
%R 10.18653/v1/2025.findings-naacl.315
%U https://aclanthology.org/2025.findings-naacl.315/
%U https://doi.org/10.18653/v1/2025.findings-naacl.315
%P 5699-5712

Markdown (Informal)

[Preserving Zero-shot Capability in Supervised Fine-tuning for Multi-label Text Classification](https://aclanthology.org/2025.findings-naacl.315/) (Chen et al., Findings 2025)

• Preserving Zero-shot Capability in Supervised Fine-tuning for Multi-label Text Classification (Chen et al., Findings 2025)

ACL

• Si-An Chen, Hsuan-Tien Lin, and Chih-Jen Lin. 2025. Preserving Zero-shot Capability in Supervised Fine-tuning for Multi-label Text Classification. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 5699–5712, Albuquerque, New Mexico. Association for Computational Linguistics.