<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-etal-2026-negation">
    <titleInfo>
        <title>Negation Without Confusion: Decomposing Bias and Recomposing Style in Text Embeddings for Negation-Aware Image Generation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Sooyeon</namePart>
        <namePart type="family">Park</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hemraj</namePart>
        <namePart type="family">Singh</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kyunghyun</namePart>
        <namePart type="family">Min</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Youngjae</namePart>
        <namePart type="family">Yu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sung-Bae</namePart>
        <namePart type="family">Cho</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Negation-aware image generation remains challenging for text-to-image diffusion models, which often struggle with negation due to implicit linguistic biases that are deeply entangled within text representations. As a result, bias-aligned semantics frequently dominate over intended negative constraints. To overcome this limitation, we propose a simple training-free method of Decompose and RecOmpose text embedding (DeRO), which constructs a semantic-preserving subspace and a bias subspace using auxiliary prompts and performs semantic disentanglement based on bidirectional SVD. By decomposing entangled representations and selectively recomposing disentangled semantic components, DeRO realigns text representations toward intended negation semantics while restoring residual stylistic features to preserve semantic fidelity and image quality. As a model-agnostic inference-time framework operating solely on text representations, DeRO can be seamlessly applied to diverse diffusion architectures, including both U-Net and DiT based models. Extensive experiments on both object- and concept-level negation demonstrate consistent improvements in negation-aware image generation, achieving stronger text-image alignment, better semantic fidelity, and higher human preference scores.</abstract>
    <identifier type="citekey">park-etal-2026-negation</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.2219/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>50492</start>
            <end>50517</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Negation Without Confusion: Decomposing Bias and Recomposing Style in Text Embeddings for Negation-Aware Image Generation
%A Park, Sooyeon
%A Singh, Hemraj
%A Min, Kyunghyun
%A Yu, Youngjae
%A Cho, Sung-Bae
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F park-etal-2026-negation
%X Negation-aware image generation remains challenging for text-to-image diffusion models, which often struggle with negation due to implicit linguistic biases that are deeply entangled within text representations. As a result, bias-aligned semantics frequently dominate over intended negative constraints. To overcome this limitation, we propose a simple training-free method of Decompose and RecOmpose text embedding (DeRO), which constructs a semantic-preserving subspace and a bias subspace using auxiliary prompts and performs semantic disentanglement based on bidirectional SVD. By decomposing entangled representations and selectively recomposing disentangled semantic components, DeRO realigns text representations toward intended negation semantics while restoring residual stylistic features to preserve semantic fidelity and image quality. As a model-agnostic inference-time framework operating solely on text representations, DeRO can be seamlessly applied to diverse diffusion architectures, including both U-Net and DiT based models. Extensive experiments on both object- and concept-level negation demonstrate consistent improvements in negation-aware image generation, achieving stronger text-image alignment, better semantic fidelity, and higher human preference scores.
%U https://aclanthology.org/2026.emnlp-main.2219/
%P 50492-50517

Markdown (Informal)

[Negation Without Confusion: Decomposing Bias and Recomposing Style in Text Embeddings for Negation-Aware Image Generation](https://aclanthology.org/2026.emnlp-main.2219/) (Park et al., EMNLP 2026)

• Negation Without Confusion: Decomposing Bias and Recomposing Style in Text Embeddings for Negation-Aware Image Generation (Park et al., EMNLP 2026)

ACL

• Sooyeon Park, Hemraj Singh, Kyunghyun Min, Youngjae Yu, and Sung-Bae Cho. 2026. Negation Without Confusion: Decomposing Bias and Recomposing Style in Text Embeddings for Negation-Aware Image Generation. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 50492–50517, Budapest, Hungary. Association for Computational Linguistics.