@inproceedings{wu-etal-2026-medocvl,
title = "{M}e{D}oc{VL}: Noise-Aware Post-Training of Visual Language Models for Medical Document Parsing",
author = "Wu, Wei and
Diao, Liang and
Zhao, Yuan and
Wang, Wenjie and
Liu, Ying and
Lv, Xiaole",
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.2321/",
pages = "49837--49859",
ISBN = "979-8-89176-462-0",
abstract = "High-precision document parsing in domain-specific scenarios, such as medical documents, remains challenging due to complex layouts, specialized terminology, and strict field-level exact-match requirements under noisy supervision. As a result, existing OCR pipelines and general-purpose vision{--}language models (VLMs) often fail to deliver reliable key{--}value extraction in practice. We propose a noise-aware post-training framework that adapts general-purpose VLMs for high-precision document parsing without assuming clean annotations. The framework learns robust extraction behavior from noisy supervision through training-driven label refinement and a staged RL-to-SFT post-training strategy that balances robustness and precision. Experiments demonstrate that VLMs adapted with the proposed framework consistently outperform strong baselines under noisy supervision."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2026-medocvl">
<titleInfo>
<title>MeDocVL: Noise-Aware Post-Training of Visual Language Models for Medical Document Parsing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liang</namePart>
<namePart type="family">Diao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuan</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenjie</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ying</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaole</namePart>
<namePart type="family">Lv</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>High-precision document parsing in domain-specific scenarios, such as medical documents, remains challenging due to complex layouts, specialized terminology, and strict field-level exact-match requirements under noisy supervision. As a result, existing OCR pipelines and general-purpose vision–language models (VLMs) often fail to deliver reliable key–value extraction in practice. We propose a noise-aware post-training framework that adapts general-purpose VLMs for high-precision document parsing without assuming clean annotations. The framework learns robust extraction behavior from noisy supervision through training-driven label refinement and a staged RL-to-SFT post-training strategy that balances robustness and precision. Experiments demonstrate that VLMs adapted with the proposed framework consistently outperform strong baselines under noisy supervision.</abstract>
<identifier type="citekey">wu-etal-2026-medocvl</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2321/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>49837</start>
<end>49859</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MeDocVL: Noise-Aware Post-Training of Visual Language Models for Medical Document Parsing
%A Wu, Wei
%A Diao, Liang
%A Zhao, Yuan
%A Wang, Wenjie
%A Liu, Ying
%A Lv, Xiaole
%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 wu-etal-2026-medocvl
%X High-precision document parsing in domain-specific scenarios, such as medical documents, remains challenging due to complex layouts, specialized terminology, and strict field-level exact-match requirements under noisy supervision. As a result, existing OCR pipelines and general-purpose vision–language models (VLMs) often fail to deliver reliable key–value extraction in practice. We propose a noise-aware post-training framework that adapts general-purpose VLMs for high-precision document parsing without assuming clean annotations. The framework learns robust extraction behavior from noisy supervision through training-driven label refinement and a staged RL-to-SFT post-training strategy that balances robustness and precision. Experiments demonstrate that VLMs adapted with the proposed framework consistently outperform strong baselines under noisy supervision.
%U https://aclanthology.org/2026.findings-emnlp.2321/
%P 49837-49859
Markdown (Informal)
[MeDocVL: Noise-Aware Post-Training of Visual Language Models for Medical Document Parsing](https://aclanthology.org/2026.findings-emnlp.2321/) (Wu et al., Findings 2026)
ACL