@inproceedings{peng-etal-2026-unidoc,
title = "{U}ni{D}oc-Bench: A Unified Benchmark for Document-Centric Multimodal {RAG}",
author = "Peng, Xiangyu and
Qin, Can and
Chen, Zeyuan and
Xu, Ran and
Xiong, Caiming and
Wu, Chien-Sheng",
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.534/",
pages = "11198--11239",
ISBN = "979-8-89176-462-0",
abstract = "Multimodal retrieval-augmented Generation (MM-RAG) is a key approach for applying large language models (LLMs) and agents to real-world knowledge bases, yet current evaluations are fragmented{---}focusing on either text or images in isolation, or simplified multimodal setup, failing to capture document-centric multimodal use cases. In this paper, we introduce UNIDOC-BENCH, the first large-scale, realistic benchmark for MM-RAG built from 70k real-world PDF pages across 8 domains. Our pipeline extracts and links evidence from text, tables, and figures, then generates 1,600 multimodal QA pairs spanning factual retrieval, comparison, summarization, and logical reasoning queries. To ensure reliability, 20{\%} of QA pairs are validated by multiple annotators and expert adjudication. UNIDOC-BENCH supports apples-to-apples comparison across four paradigms {---} 1) text-only, 2) image-only, 3) multimodal text{--}image fusion and 4) multimodal joint retrieval {---} under a unified protocol with standardized candidate pools, prompts, and evaluation metrics. Our experiments show that multimodal text{--}image fusion RAG systems consistently outperform both unimodal and jointly multimodal embedding{--}based retrieval, indicating that neither text nor images alone are sufficient and that current multimodal embeddings remain inadequate. Beyond benchmarking, our analysis reveals when and how visual context complements textual evidence, uncovers systematic failure modes, and offers actionable guidance for developing more robust MM-RAG pipelines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="peng-etal-2026-unidoc">
<titleInfo>
<title>UniDoc-Bench: A Unified Benchmark for Document-Centric Multimodal RAG</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xiangyu</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Can</namePart>
<namePart type="family">Qin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zeyuan</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ran</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Caiming</namePart>
<namePart type="family">Xiong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chien-Sheng</namePart>
<namePart type="family">Wu</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>Multimodal retrieval-augmented Generation (MM-RAG) is a key approach for applying large language models (LLMs) and agents to real-world knowledge bases, yet current evaluations are fragmented—focusing on either text or images in isolation, or simplified multimodal setup, failing to capture document-centric multimodal use cases. In this paper, we introduce UNIDOC-BENCH, the first large-scale, realistic benchmark for MM-RAG built from 70k real-world PDF pages across 8 domains. Our pipeline extracts and links evidence from text, tables, and figures, then generates 1,600 multimodal QA pairs spanning factual retrieval, comparison, summarization, and logical reasoning queries. To ensure reliability, 20% of QA pairs are validated by multiple annotators and expert adjudication. UNIDOC-BENCH supports apples-to-apples comparison across four paradigms — 1) text-only, 2) image-only, 3) multimodal text–image fusion and 4) multimodal joint retrieval — under a unified protocol with standardized candidate pools, prompts, and evaluation metrics. Our experiments show that multimodal text–image fusion RAG systems consistently outperform both unimodal and jointly multimodal embedding–based retrieval, indicating that neither text nor images alone are sufficient and that current multimodal embeddings remain inadequate. Beyond benchmarking, our analysis reveals when and how visual context complements textual evidence, uncovers systematic failure modes, and offers actionable guidance for developing more robust MM-RAG pipelines.</abstract>
<identifier type="citekey">peng-etal-2026-unidoc</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.534/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>11198</start>
<end>11239</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T UniDoc-Bench: A Unified Benchmark for Document-Centric Multimodal RAG
%A Peng, Xiangyu
%A Qin, Can
%A Chen, Zeyuan
%A Xu, Ran
%A Xiong, Caiming
%A Wu, Chien-Sheng
%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 peng-etal-2026-unidoc
%X Multimodal retrieval-augmented Generation (MM-RAG) is a key approach for applying large language models (LLMs) and agents to real-world knowledge bases, yet current evaluations are fragmented—focusing on either text or images in isolation, or simplified multimodal setup, failing to capture document-centric multimodal use cases. In this paper, we introduce UNIDOC-BENCH, the first large-scale, realistic benchmark for MM-RAG built from 70k real-world PDF pages across 8 domains. Our pipeline extracts and links evidence from text, tables, and figures, then generates 1,600 multimodal QA pairs spanning factual retrieval, comparison, summarization, and logical reasoning queries. To ensure reliability, 20% of QA pairs are validated by multiple annotators and expert adjudication. UNIDOC-BENCH supports apples-to-apples comparison across four paradigms — 1) text-only, 2) image-only, 3) multimodal text–image fusion and 4) multimodal joint retrieval — under a unified protocol with standardized candidate pools, prompts, and evaluation metrics. Our experiments show that multimodal text–image fusion RAG systems consistently outperform both unimodal and jointly multimodal embedding–based retrieval, indicating that neither text nor images alone are sufficient and that current multimodal embeddings remain inadequate. Beyond benchmarking, our analysis reveals when and how visual context complements textual evidence, uncovers systematic failure modes, and offers actionable guidance for developing more robust MM-RAG pipelines.
%U https://aclanthology.org/2026.findings-emnlp.534/
%P 11198-11239
Markdown (Informal)
[UniDoc-Bench: A Unified Benchmark for Document-Centric Multimodal RAG](https://aclanthology.org/2026.findings-emnlp.534/) (Peng et al., Findings 2026)
ACL