@inproceedings{zhao-etal-2026-globe,
title = "{GLOBE}: Training-Free Calibration for Image-Text Retrieval with Global-Local Bias and Graph Evidence",
author = "Zhao, Xuzhe and
Yue, Zhirun and
Xu, Xingyue and
Liu, Houde and
Dong, Kaichen",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1985/",
pages = "45154--45173",
ISBN = "979-8-89176-459-0",
abstract = "Multimodal contrastive models are widely used for bidirectional image-text retrieval, but their rankings remain vulnerable to attraction bias, where certain candidates become hubs for many unrelated queries. Existing training-free normalization methods suppress such hubs, but may also penalize candidates that are genuinely relevant. We propose GLOBE, a training-free plug-in module that combines global-local bias calibration with graph-based semantic evidence to suppress unsupported hubs while preserving true matches. Across MS-COCO and Flickr30K, GLOBE improves diverse frozen retrieval backbones, achieving up to 7.9{\%} absolute Recall@1 gain over the original model and up to 5.2{\%} over strong training-free baselines. We further construct IRB-COCO-500, a mixed real-and-synthetic benchmark with AI-generated near-positive distractors, and show that GLOBE improves retrieval robustness in this challenging setting."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhao-etal-2026-globe">
<titleInfo>
<title>GLOBE: Training-Free Calibration for Image-Text Retrieval with Global-Local Bias and Graph Evidence</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xuzhe</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhirun</namePart>
<namePart type="family">Yue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xingyue</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Houde</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kaichen</namePart>
<namePart type="family">Dong</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>Multimodal contrastive models are widely used for bidirectional image-text retrieval, but their rankings remain vulnerable to attraction bias, where certain candidates become hubs for many unrelated queries. Existing training-free normalization methods suppress such hubs, but may also penalize candidates that are genuinely relevant. We propose GLOBE, a training-free plug-in module that combines global-local bias calibration with graph-based semantic evidence to suppress unsupported hubs while preserving true matches. Across MS-COCO and Flickr30K, GLOBE improves diverse frozen retrieval backbones, achieving up to 7.9% absolute Recall@1 gain over the original model and up to 5.2% over strong training-free baselines. We further construct IRB-COCO-500, a mixed real-and-synthetic benchmark with AI-generated near-positive distractors, and show that GLOBE improves retrieval robustness in this challenging setting.</abstract>
<identifier type="citekey">zhao-etal-2026-globe</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1985/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>45154</start>
<end>45173</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GLOBE: Training-Free Calibration for Image-Text Retrieval with Global-Local Bias and Graph Evidence
%A Zhao, Xuzhe
%A Yue, Zhirun
%A Xu, Xingyue
%A Liu, Houde
%A Dong, Kaichen
%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 zhao-etal-2026-globe
%X Multimodal contrastive models are widely used for bidirectional image-text retrieval, but their rankings remain vulnerable to attraction bias, where certain candidates become hubs for many unrelated queries. Existing training-free normalization methods suppress such hubs, but may also penalize candidates that are genuinely relevant. We propose GLOBE, a training-free plug-in module that combines global-local bias calibration with graph-based semantic evidence to suppress unsupported hubs while preserving true matches. Across MS-COCO and Flickr30K, GLOBE improves diverse frozen retrieval backbones, achieving up to 7.9% absolute Recall@1 gain over the original model and up to 5.2% over strong training-free baselines. We further construct IRB-COCO-500, a mixed real-and-synthetic benchmark with AI-generated near-positive distractors, and show that GLOBE improves retrieval robustness in this challenging setting.
%U https://aclanthology.org/2026.emnlp-main.1985/
%P 45154-45173
Markdown (Informal)
[GLOBE: Training-Free Calibration for Image-Text Retrieval with Global-Local Bias and Graph Evidence](https://aclanthology.org/2026.emnlp-main.1985/) (Zhao et al., EMNLP 2026)
ACL