<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kao-etal-2025-qg">
    <titleInfo>
        <title>QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Kuei-Chun</namePart>
        <namePart type="family">Kao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hsu</namePart>
        <namePart type="family">Tzu-Yin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yunqi</namePart>
        <namePart type="family">Hong</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ruochen</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Cho-Jui</namePart>
        <namePart type="family">Hsieh</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>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-332-6</identifier>
    </relatedItem>
    <abstract>Recently, Multimodal Large Language Models (MLLMs) encounter two key issues in multi-image contexts: (1) a lack of fine-grained perception across disparate images, and (2) a diminished capability to effectively reason over and synthesize information from multiple visual inputs. However, while various prompting methods aim to describe visual content, many existing studies focus primarily on single-image settings or specific, constrained scenarios. This leaves a critical gap in understanding and addressing how MLLMs tackle more general and complex multi-image reasoning tasks. Thus, we first extensively investigate how current prompting methods perceive fine-grained visual details and process visual information when dealing with multiple images. Our findings reveal that existing prompting methods fall short in attending to needed clues and seamlessly integrating perception and reasoning. Inspired by the findings, we propose a new zero-shot prompting method, Question-Guided Chain-of-Captions (QG-CoC), a generalized prompting approach that effectively handles problems with an arbitrary number of images. We evaluate our method on various open-source and closed-source MLLMs for multi-image and single-image benchmarks. Experimental results indicate that QG-CoC demonstrates competitive performance across tasks and exhibits robust improvements in the challenging scenarios where existing prompting methods fail.</abstract>
    <identifier type="citekey">kao-etal-2025-qg</identifier>
    <identifier type="doi">10.18653/v1/2025.emnlp-main.1445</identifier>
    <location>
        <url>https://aclanthology.org/2025.emnlp-main.1445/</url>
    </location>
    <part>
        <date>2025-11</date>
        <extent unit="page">
            <start>28445</start>
            <end>28460</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models
%A Kao, Kuei-Chun
%A Tzu-Yin, Hsu
%A Hong, Yunqi
%A Wang, Ruochen
%A Hsieh, Cho-Jui
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F kao-etal-2025-qg
%X Recently, Multimodal Large Language Models (MLLMs) encounter two key issues in multi-image contexts: (1) a lack of fine-grained perception across disparate images, and (2) a diminished capability to effectively reason over and synthesize information from multiple visual inputs. However, while various prompting methods aim to describe visual content, many existing studies focus primarily on single-image settings or specific, constrained scenarios. This leaves a critical gap in understanding and addressing how MLLMs tackle more general and complex multi-image reasoning tasks. Thus, we first extensively investigate how current prompting methods perceive fine-grained visual details and process visual information when dealing with multiple images. Our findings reveal that existing prompting methods fall short in attending to needed clues and seamlessly integrating perception and reasoning. Inspired by the findings, we propose a new zero-shot prompting method, Question-Guided Chain-of-Captions (QG-CoC), a generalized prompting approach that effectively handles problems with an arbitrary number of images. We evaluate our method on various open-source and closed-source MLLMs for multi-image and single-image benchmarks. Experimental results indicate that QG-CoC demonstrates competitive performance across tasks and exhibits robust improvements in the challenging scenarios where existing prompting methods fail.
%R 10.18653/v1/2025.emnlp-main.1445
%U https://aclanthology.org/2025.emnlp-main.1445/
%U https://doi.org/10.18653/v1/2025.emnlp-main.1445
%P 28445-28460

Markdown (Informal)

[QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models](https://aclanthology.org/2025.emnlp-main.1445/) (Kao et al., EMNLP 2025)

• QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models (Kao et al., EMNLP 2025)

ACL

• Kuei-Chun Kao, Hsu Tzu-Yin, Yunqi Hong, Ruochen Wang, and Cho-Jui Hsieh. 2025. QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 28445–28460, Suzhou, China. Association for Computational Linguistics.