@inproceedings{wang-etal-2026-seeing-agreeing,
title = "Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence",
author = "Wang, Yuhan and
Chang, Shuochen and
Feng, Yalin and
Ma, Dongsheng and
Li, Yuanzi and
Wang, Zhengren and
Yang, Yinglong and
Chen, Yufei and
Wang, Yikang and
Sun, Shaoxu and
Zhang, Wentao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.473/",
pages = "9925--9952",
ISBN = "979-8-89176-462-0",
abstract = "Vision-language models (VLMs) have achieved strong performance on visual question answering (VQA). To mitigate individual hallucinations and blind spots, aggregating diverse perspectives via multi-agent collaboration has emerged as a promising paradigm. While this approach has shown great success in textual QA, its potential in the multimodal domain remains under-explored. Existing multi-agent VQA methods predominantly adapt text-centric protocols, focusing on textual discussions while ignoring the alignment of visual information. In this work, we reveal a key insight: answer-level agreement is insufficient for reliable multi-agent VQA; *aligned visual evidence*{---}shared support from the image regions agents rely on{---}is essential for trustworthy consensus. To leverage this insight, we propose **EAGLE** (**E**vidence-**A**ligned **G**rounded mu**L**ti-agent r**E**asoning), a training-free evidence-centered framework for coordinating multiple VLM agents. **EAGLE** explicitly exposes each agent{'}s grounding regions as visual evidence, enables mutual verification over the evidence, and uses evidence consistency to guide final decision-making. Experiments on six VQA benchmarks show that **EAGLE** achieves best average performance across domains while remaining lightweight, interpretable, and practical for deployment."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-seeing-agreeing">
<titleInfo>
<title>Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuhan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuochen</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yalin</namePart>
<namePart type="family">Feng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dongsheng</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuanzi</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengren</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yinglong</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yufei</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yikang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaoxu</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wentao</namePart>
<namePart type="family">Zhang</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Vision-language models (VLMs) have achieved strong performance on visual question answering (VQA). To mitigate individual hallucinations and blind spots, aggregating diverse perspectives via multi-agent collaboration has emerged as a promising paradigm. While this approach has shown great success in textual QA, its potential in the multimodal domain remains under-explored. Existing multi-agent VQA methods predominantly adapt text-centric protocols, focusing on textual discussions while ignoring the alignment of visual information. In this work, we reveal a key insight: answer-level agreement is insufficient for reliable multi-agent VQA; *aligned visual evidence*—shared support from the image regions agents rely on—is essential for trustworthy consensus. To leverage this insight, we propose **EAGLE** (**E**vidence-**A**ligned **G**rounded mu**L**ti-agent r**E**asoning), a training-free evidence-centered framework for coordinating multiple VLM agents. **EAGLE** explicitly exposes each agent’s grounding regions as visual evidence, enables mutual verification over the evidence, and uses evidence consistency to guide final decision-making. Experiments on six VQA benchmarks show that **EAGLE** achieves best average performance across domains while remaining lightweight, interpretable, and practical for deployment.</abstract>
<identifier type="citekey">wang-etal-2026-seeing-agreeing</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.473/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>9925</start>
<end>9952</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence
%A Wang, Yuhan
%A Chang, Shuochen
%A Feng, Yalin
%A Ma, Dongsheng
%A Li, Yuanzi
%A Wang, Zhengren
%A Yang, Yinglong
%A Chen, Yufei
%A Wang, Yikang
%A Sun, Shaoxu
%A Zhang, Wentao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F wang-etal-2026-seeing-agreeing
%X Vision-language models (VLMs) have achieved strong performance on visual question answering (VQA). To mitigate individual hallucinations and blind spots, aggregating diverse perspectives via multi-agent collaboration has emerged as a promising paradigm. While this approach has shown great success in textual QA, its potential in the multimodal domain remains under-explored. Existing multi-agent VQA methods predominantly adapt text-centric protocols, focusing on textual discussions while ignoring the alignment of visual information. In this work, we reveal a key insight: answer-level agreement is insufficient for reliable multi-agent VQA; *aligned visual evidence*—shared support from the image regions agents rely on—is essential for trustworthy consensus. To leverage this insight, we propose **EAGLE** (**E**vidence-**A**ligned **G**rounded mu**L**ti-agent r**E**asoning), a training-free evidence-centered framework for coordinating multiple VLM agents. **EAGLE** explicitly exposes each agent’s grounding regions as visual evidence, enables mutual verification over the evidence, and uses evidence consistency to guide final decision-making. Experiments on six VQA benchmarks show that **EAGLE** achieves best average performance across domains while remaining lightweight, interpretable, and practical for deployment.
%U https://aclanthology.org/2026.findings-emnlp.473/
%P 9925-9952
Markdown (Informal)
[Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence](https://aclanthology.org/2026.findings-emnlp.473/) (Wang et al., Findings 2026)
ACL
- Yuhan Wang, Shuochen Chang, Yalin Feng, Dongsheng Ma, Yuanzi Li, Zhengren Wang, Yinglong Yang, Yufei Chen, Yikang Wang, Shaoxu Sun, and Wentao Zhang. 2026. Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 9925–9952, Budapest, Hungary. Association for Computational Linguistics.