@inproceedings{li-etal-2026-llava,
title = "{LL}a{VA}-{R}ad{Z}: Can Multimodal Large Language Models Effectively Tackle Zero-shot Radiology Recognition?",
author = "Li, Bangyan and
Huang, Wenxuan and
Gao, Zhenkun and
Wang, Yeqiang and
Shen, Yuxiang",
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.1204/",
pages = "25435--25454",
ISBN = "979-8-89176-462-0",
abstract = "Recently, Multimodal Large Language Models (MLLMs) have demonstrated strong visual understanding and reasoning abilities across vision-language tasks. However, conventional Visual Question Answering (VQA) pipelines do not fully exploit the visual features and medical knowledge captured by MLLMs, leading to limited performance in zero-shot medical disease recognition. To address this issue, we propose LLaVA-RadZ, a simple yet effective framework for zero-shot radiology recognition. LLaVA-RadZ introduces Decoding-Side Feature Alignment Training (DFAT), which leverages decoder-side representations with modality-specific tokens for image-text alignment, and a Domain Knowledge Anchoring Module (DKAM), which incorporates disease-aware semantic anchors to mitigate the category semantic gap. Extensive experiments show that LLaVA-RadZ substantially improves the zero-shot disease recognition ability of MLLMs and achieves competitive performance compared with well-established CLIP-based radiology recognition methods."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-llava">
<titleInfo>
<title>LLaVA-RadZ: Can Multimodal Large Language Models Effectively Tackle Zero-shot Radiology Recognition?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bangyan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenxuan</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenkun</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yeqiang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuxiang</namePart>
<namePart type="family">Shen</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>Recently, Multimodal Large Language Models (MLLMs) have demonstrated strong visual understanding and reasoning abilities across vision-language tasks. However, conventional Visual Question Answering (VQA) pipelines do not fully exploit the visual features and medical knowledge captured by MLLMs, leading to limited performance in zero-shot medical disease recognition. To address this issue, we propose LLaVA-RadZ, a simple yet effective framework for zero-shot radiology recognition. LLaVA-RadZ introduces Decoding-Side Feature Alignment Training (DFAT), which leverages decoder-side representations with modality-specific tokens for image-text alignment, and a Domain Knowledge Anchoring Module (DKAM), which incorporates disease-aware semantic anchors to mitigate the category semantic gap. Extensive experiments show that LLaVA-RadZ substantially improves the zero-shot disease recognition ability of MLLMs and achieves competitive performance compared with well-established CLIP-based radiology recognition methods.</abstract>
<identifier type="citekey">li-etal-2026-llava</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1204/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>25435</start>
<end>25454</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LLaVA-RadZ: Can Multimodal Large Language Models Effectively Tackle Zero-shot Radiology Recognition?
%A Li, Bangyan
%A Huang, Wenxuan
%A Gao, Zhenkun
%A Wang, Yeqiang
%A Shen, Yuxiang
%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 li-etal-2026-llava
%X Recently, Multimodal Large Language Models (MLLMs) have demonstrated strong visual understanding and reasoning abilities across vision-language tasks. However, conventional Visual Question Answering (VQA) pipelines do not fully exploit the visual features and medical knowledge captured by MLLMs, leading to limited performance in zero-shot medical disease recognition. To address this issue, we propose LLaVA-RadZ, a simple yet effective framework for zero-shot radiology recognition. LLaVA-RadZ introduces Decoding-Side Feature Alignment Training (DFAT), which leverages decoder-side representations with modality-specific tokens for image-text alignment, and a Domain Knowledge Anchoring Module (DKAM), which incorporates disease-aware semantic anchors to mitigate the category semantic gap. Extensive experiments show that LLaVA-RadZ substantially improves the zero-shot disease recognition ability of MLLMs and achieves competitive performance compared with well-established CLIP-based radiology recognition methods.
%U https://aclanthology.org/2026.findings-emnlp.1204/
%P 25435-25454
Markdown (Informal)
[LLaVA-RadZ: Can Multimodal Large Language Models Effectively Tackle Zero-shot Radiology Recognition?](https://aclanthology.org/2026.findings-emnlp.1204/) (Li et al., Findings 2026)
ACL