<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-anonymous">
    <titleInfo>
        <title>From Anonymous Speaker Seeds to Consistent Characters: Zero-Shot Manga Speaker Attribution</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yanhao</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Wataru</namePart>
        <namePart type="family">Shimoda</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kota</namePart>
        <namePart type="family">Yamaguchi</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hideki</namePart>
        <namePart type="family">Nakayama</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>Identifying the speaker of each dialogue instance is a key step in manga understanding, supporting applications such as translation, summarization, and voice generation. In zero-shot settings, the task is challenging because each new manga title introduces an unseen cast and provides no speaker annotations, predefined character identities, or per-title model adaptation. Existing methods, including iterative multimodal fusion, still organize speaker reasoning around manga-specific visual evidence such as characters’ appearance and local text-character relationships. However, such evidence is highly variable in manga, where speakers may be off-panel, character appearance changes across poses and styles, and local visual relations can be misleading. We propose Speaker-Seeded Identity Consolidation (SSIC), which reframes zero-shot manga speaker attribution as book-level identity consolidation driven by recurring dialogue consistency rather than local visual-region propagation. SSIC induces local anonymous speaker hypotheses from ordered dialogue context, aligns them into cross-window speaker seeds, constructs a global speaker identity registry, and performs registry-constrained attribution. In the main controlled evaluation on Manga109Dialog, SSIC achieves 74.62% speaker attribution accuracy, improving over iterative multimodal fusion by 7.02 points. The gains are largest on hard cases where local text-character relationships are unreliable, and ablations confirm that cross-window speaker seeds are central to registry construction.</abstract>
    <identifier type="citekey">chen-etal-2026-anonymous</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.2632/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>59842</start>
            <end>59855</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T From Anonymous Speaker Seeds to Consistent Characters: Zero-Shot Manga Speaker Attribution
%A Chen, Yanhao
%A Shimoda, Wataru
%A Yamaguchi, Kota
%A Nakayama, Hideki
%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 chen-etal-2026-anonymous
%X Identifying the speaker of each dialogue instance is a key step in manga understanding, supporting applications such as translation, summarization, and voice generation. In zero-shot settings, the task is challenging because each new manga title introduces an unseen cast and provides no speaker annotations, predefined character identities, or per-title model adaptation. Existing methods, including iterative multimodal fusion, still organize speaker reasoning around manga-specific visual evidence such as characters’ appearance and local text-character relationships. However, such evidence is highly variable in manga, where speakers may be off-panel, character appearance changes across poses and styles, and local visual relations can be misleading. We propose Speaker-Seeded Identity Consolidation (SSIC), which reframes zero-shot manga speaker attribution as book-level identity consolidation driven by recurring dialogue consistency rather than local visual-region propagation. SSIC induces local anonymous speaker hypotheses from ordered dialogue context, aligns them into cross-window speaker seeds, constructs a global speaker identity registry, and performs registry-constrained attribution. In the main controlled evaluation on Manga109Dialog, SSIC achieves 74.62% speaker attribution accuracy, improving over iterative multimodal fusion by 7.02 points. The gains are largest on hard cases where local text-character relationships are unreliable, and ablations confirm that cross-window speaker seeds are central to registry construction.
%U https://aclanthology.org/2026.emnlp-main.2632/
%P 59842-59855

Markdown (Informal)

[From Anonymous Speaker Seeds to Consistent Characters: Zero-Shot Manga Speaker Attribution](https://aclanthology.org/2026.emnlp-main.2632/) (Chen et al., EMNLP 2026)

• From Anonymous Speaker Seeds to Consistent Characters: Zero-Shot Manga Speaker Attribution (Chen et al., EMNLP 2026)

ACL

• Yanhao Chen, Wataru Shimoda, Kota Yamaguchi, and Hideki Nakayama. 2026. From Anonymous Speaker Seeds to Consistent Characters: Zero-Shot Manga Speaker Attribution. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 59842–59855, Budapest, Hungary. Association for Computational Linguistics.