<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-etal-2026-mitigating">
    <titleInfo>
        <title>Mitigating Cross-Image Information Leakage in Multi-Image Understanding with Large Vision-Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yeji</namePart>
        <namePart type="family">Park</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Minyoung</namePart>
        <namePart type="family">Lee</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sanghyuk</namePart>
        <namePart type="family">Chun</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Junsuk</namePart>
        <namePart type="family">Choe</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yue</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ruifeng</namePart>
            <namePart type="family">Xu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ayu</namePart>
            <namePart type="family">Purwarianti</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tong</namePart>
            <namePart type="family">Rong</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jun</namePart>
            <namePart type="family">Suzuki</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Hengqin, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-440-8</identifier>
    </relatedItem>
    <abstract>Large Vision-Language Models (LVLMs) exhibit strong performance on single-image tasks. However, their performance degrades significantly when handling multi-image inputs. While this degradation has been observed in prior work, its nature remains poorly understood. We empirically observe visual elements from different images become entangled in the model’s representations and responses. We refer to this phenomenon as cross-image information leakage. To address this issue, we propose FOCUS, a training-free and architecture-agnostic method. FOCUS masks all but one image with random noise, guiding the model to focus on the single clean image. This process is applied across the target images to obtain logits under partially masked contexts. These logits are aggregated and then refined using a noise-only reference input, which suppresses the leakage and yields more accurate outputs. FOCUS consistently improves performance on diverse multi-image benchmarks. We further show that generalizes to video understanding, extending its applicability beyond static multi-image inputs. This demonstrates that FOCUS offers a general solution for enhancing multi-image reasoning without additional training or architectural modifications.</abstract>
    <identifier type="citekey">park-etal-2026-mitigating</identifier>
    <location>
        <url>https://aclanthology.org/2026.aacl-short.42/</url>
    </location>
    <part>
        <date>2026-11</date>
        <extent unit="page">
            <start>645</start>
            <end>663</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Mitigating Cross-Image Information Leakage in Multi-Image Understanding with Large Vision-Language Models
%A Park, Yeji
%A Lee, Minyoung
%A Chun, Sanghyuk
%A Choe, Junsuk
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-440-8
%F park-etal-2026-mitigating
%X Large Vision-Language Models (LVLMs) exhibit strong performance on single-image tasks. However, their performance degrades significantly when handling multi-image inputs. While this degradation has been observed in prior work, its nature remains poorly understood. We empirically observe visual elements from different images become entangled in the model’s representations and responses. We refer to this phenomenon as cross-image information leakage. To address this issue, we propose FOCUS, a training-free and architecture-agnostic method. FOCUS masks all but one image with random noise, guiding the model to focus on the single clean image. This process is applied across the target images to obtain logits under partially masked contexts. These logits are aggregated and then refined using a noise-only reference input, which suppresses the leakage and yields more accurate outputs. FOCUS consistently improves performance on diverse multi-image benchmarks. We further show that generalizes to video understanding, extending its applicability beyond static multi-image inputs. This demonstrates that FOCUS offers a general solution for enhancing multi-image reasoning without additional training or architectural modifications.
%U https://aclanthology.org/2026.aacl-short.42/
%P 645-663

Markdown (Informal)

[Mitigating Cross-Image Information Leakage in Multi-Image Understanding with Large Vision-Language Models](https://aclanthology.org/2026.aacl-short.42/) (Park et al., AACL-IJCNLP 2026)

• Mitigating Cross-Image Information Leakage in Multi-Image Understanding with Large Vision-Language Models (Park et al., AACL-IJCNLP 2026)

ACL

• Yeji Park, Minyoung Lee, Sanghyuk Chun, and Junsuk Choe. 2026. Mitigating Cross-Image Information Leakage in Multi-Image Understanding with Large Vision-Language Models. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers), pages 645–663, Hengqin, China. Association for Computational Linguistics.