@inproceedings{gangavarapu-2026-hgkai26,
title = "hgkai26 at {MEDIQA}-{EVAL} 2026: Automated Evaluation of Visual Medical Question Answering Using {LLM}-as-a-Judge",
author = "Gangavarapu, Haritha",
editor = "Ben Abacha, Asma and
Bethard, Steven and
Bitterman, Danielle and
Naumann, Tristan and
Roberts, Kirk",
booktitle = "Proceedings of the 8th Workshop on Clinical Natural Language Processing (Clinical {NLP}) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.clinicalnlp-1.29/",
doi = "10.63317/4n9skmf9rive",
pages = "257--261",
abstract = "As there is a rise in the use of multimodal large language models (LLMs) for medical response generation, it is necessary to have reliable automated evaluation mechanisms that can assess the quality of model-generated outputs. The MediQA-Eval 2026 shared task focuses on grading AI-generated dermatology and wound care responses using structured human-aligned rubrics. In this work, we explore a zero-shot multimodal LLM-as-a-Judge framework to assess candidate responses across multiple quality dimensions. System performance is evaluated using the official task metrics designed to reflect alignment with human judgments. Our findings provide preliminary insights into the feasibility and limitations of LLM-based evaluators for rubric-guided medical response assessment."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="gangavarapu-2026-hgkai26">
<titleInfo>
<title>hgkai26 at MEDIQA-EVAL 2026: Automated Evaluation of Visual Medical Question Answering Using LLM-as-a-Judge</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haritha</namePart>
<namePart type="family">Gangavarapu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 8th Workshop on Clinical Natural Language Processing (Clinical NLP) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Asma</namePart>
<namePart type="family">Ben Abacha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Bethard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Danielle</namePart>
<namePart type="family">Bitterman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tristan</namePart>
<namePart type="family">Naumann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kirk</namePart>
<namePart type="family">Roberts</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>As there is a rise in the use of multimodal large language models (LLMs) for medical response generation, it is necessary to have reliable automated evaluation mechanisms that can assess the quality of model-generated outputs. The MediQA-Eval 2026 shared task focuses on grading AI-generated dermatology and wound care responses using structured human-aligned rubrics. In this work, we explore a zero-shot multimodal LLM-as-a-Judge framework to assess candidate responses across multiple quality dimensions. System performance is evaluated using the official task metrics designed to reflect alignment with human judgments. Our findings provide preliminary insights into the feasibility and limitations of LLM-based evaluators for rubric-guided medical response assessment.</abstract>
<identifier type="citekey">gangavarapu-2026-hgkai26</identifier>
<identifier type="doi">10.63317/4n9skmf9rive</identifier>
<location>
<url>https://aclanthology.org/2026.clinicalnlp-1.29/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>257</start>
<end>261</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T hgkai26 at MEDIQA-EVAL 2026: Automated Evaluation of Visual Medical Question Answering Using LLM-as-a-Judge
%A Gangavarapu, Haritha
%Y Ben Abacha, Asma
%Y Bethard, Steven
%Y Bitterman, Danielle
%Y Naumann, Tristan
%Y Roberts, Kirk
%S Proceedings of the 8th Workshop on Clinical Natural Language Processing (Clinical NLP) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F gangavarapu-2026-hgkai26
%X As there is a rise in the use of multimodal large language models (LLMs) for medical response generation, it is necessary to have reliable automated evaluation mechanisms that can assess the quality of model-generated outputs. The MediQA-Eval 2026 shared task focuses on grading AI-generated dermatology and wound care responses using structured human-aligned rubrics. In this work, we explore a zero-shot multimodal LLM-as-a-Judge framework to assess candidate responses across multiple quality dimensions. System performance is evaluated using the official task metrics designed to reflect alignment with human judgments. Our findings provide preliminary insights into the feasibility and limitations of LLM-based evaluators for rubric-guided medical response assessment.
%R 10.63317/4n9skmf9rive
%U https://aclanthology.org/2026.clinicalnlp-1.29/
%U https://doi.org/10.63317/4n9skmf9rive
%P 257-261
Markdown (Informal)
[hgkai26 at MEDIQA-EVAL 2026: Automated Evaluation of Visual Medical Question Answering Using LLM-as-a-Judge](https://aclanthology.org/2026.clinicalnlp-1.29/) (Gangavarapu, ClinicalNLP 2026)
ACL