@inproceedings{xu-etal-2026-bdi,
title = "{BDI} at {MEDIQA}-{EVAL} 2026: A {R}e{A}ct-Style Multimodal Agent for Fine-Grained Medical Response Assessment",
author = "Xu, Justin and
Zhang, Zizheng and
Luk, Augustine and
Khong, Benjamin and
Cui, Haochen and
Hwang, Samuel and
Pradhan, Alyssa and
Yuan, Kevin and
Eyre, David W.",
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-1.16/",
doi = "10.63317/3xgwaboh563b",
pages = "141--152",
abstract = "Free-text evaluation of multimodal clinical question answering (QA) systems remains a central challenge in medical NLP due to the complexity of medical knowledge, the necessity of integrating visual and textual information, and the limitations of existing automatic evaluation metrics for open-ended outputs. In this work, we present a training-free, agentic evaluation framework that formulates response scoring as evidence-guided orchestration of components rather than a task requiring conventional end-to-end fine-tuning of underlying LLMs/VLMs. Our ReAct-style evaluator combines (i) structured reasoning, (ii) multimodal retrieval of similar encounters, (iii) auxiliary explainable feature-based regression models that provide numeric priors and human-interpretable signals, (iv) VLM-generated visual QA references for comparison, and (v) optional image augmentation tools. Unlike standard LLM-as-a-judge approaches that rely on direct generative scoring, our agent decomposes evaluation into modular stages of evidence acquisition, structured feature modeling, and integrative reasoning. We apply this architecture to the MEDIQA-EVAL shared task - a multimodal, multilingual clinical evaluation challenge that assesses system-generated answers for patient queries paired with images along multiple clinical quality dimensions. We report results across both English and Chinese tracks, comparing against baseline prompting methods, and discuss the feasibility and limitations of lightweight agentic systems for clinical QA evaluation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xu-etal-2026-bdi">
<titleInfo>
<title>BDI at MEDIQA-EVAL 2026: A ReAct-Style Multimodal Agent for Fine-Grained Medical Response Assessment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Justin</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zizheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Augustine</namePart>
<namePart type="family">Luk</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Benjamin</namePart>
<namePart type="family">Khong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haochen</namePart>
<namePart type="family">Cui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samuel</namePart>
<namePart type="family">Hwang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alyssa</namePart>
<namePart type="family">Pradhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Yuan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="given">W</namePart>
<namePart type="family">Eyre</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>Free-text evaluation of multimodal clinical question answering (QA) systems remains a central challenge in medical NLP due to the complexity of medical knowledge, the necessity of integrating visual and textual information, and the limitations of existing automatic evaluation metrics for open-ended outputs. In this work, we present a training-free, agentic evaluation framework that formulates response scoring as evidence-guided orchestration of components rather than a task requiring conventional end-to-end fine-tuning of underlying LLMs/VLMs. Our ReAct-style evaluator combines (i) structured reasoning, (ii) multimodal retrieval of similar encounters, (iii) auxiliary explainable feature-based regression models that provide numeric priors and human-interpretable signals, (iv) VLM-generated visual QA references for comparison, and (v) optional image augmentation tools. Unlike standard LLM-as-a-judge approaches that rely on direct generative scoring, our agent decomposes evaluation into modular stages of evidence acquisition, structured feature modeling, and integrative reasoning. We apply this architecture to the MEDIQA-EVAL shared task - a multimodal, multilingual clinical evaluation challenge that assesses system-generated answers for patient queries paired with images along multiple clinical quality dimensions. We report results across both English and Chinese tracks, comparing against baseline prompting methods, and discuss the feasibility and limitations of lightweight agentic systems for clinical QA evaluation.</abstract>
<identifier type="citekey">xu-etal-2026-bdi</identifier>
<identifier type="doi">10.63317/3xgwaboh563b</identifier>
<location>
<url>https://aclanthology.org/2026.clinicalnlp-1.16/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>141</start>
<end>152</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T BDI at MEDIQA-EVAL 2026: A ReAct-Style Multimodal Agent for Fine-Grained Medical Response Assessment
%A Xu, Justin
%A Zhang, Zizheng
%A Luk, Augustine
%A Khong, Benjamin
%A Cui, Haochen
%A Hwang, Samuel
%A Pradhan, Alyssa
%A Yuan, Kevin
%A Eyre, David W.
%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 xu-etal-2026-bdi
%X Free-text evaluation of multimodal clinical question answering (QA) systems remains a central challenge in medical NLP due to the complexity of medical knowledge, the necessity of integrating visual and textual information, and the limitations of existing automatic evaluation metrics for open-ended outputs. In this work, we present a training-free, agentic evaluation framework that formulates response scoring as evidence-guided orchestration of components rather than a task requiring conventional end-to-end fine-tuning of underlying LLMs/VLMs. Our ReAct-style evaluator combines (i) structured reasoning, (ii) multimodal retrieval of similar encounters, (iii) auxiliary explainable feature-based regression models that provide numeric priors and human-interpretable signals, (iv) VLM-generated visual QA references for comparison, and (v) optional image augmentation tools. Unlike standard LLM-as-a-judge approaches that rely on direct generative scoring, our agent decomposes evaluation into modular stages of evidence acquisition, structured feature modeling, and integrative reasoning. We apply this architecture to the MEDIQA-EVAL shared task - a multimodal, multilingual clinical evaluation challenge that assesses system-generated answers for patient queries paired with images along multiple clinical quality dimensions. We report results across both English and Chinese tracks, comparing against baseline prompting methods, and discuss the feasibility and limitations of lightweight agentic systems for clinical QA evaluation.
%R 10.63317/3xgwaboh563b
%U https://aclanthology.org/2026.clinicalnlp-1.16/
%U https://doi.org/10.63317/3xgwaboh563b
%P 141-152
Markdown (Informal)
[BDI at MEDIQA-EVAL 2026: A ReAct-Style Multimodal Agent for Fine-Grained Medical Response Assessment](https://aclanthology.org/2026.clinicalnlp-1.16/) (Xu et al., ClinicalNLP 2026)
ACL
- Justin Xu, Zizheng Zhang, Augustine Luk, Benjamin Khong, Haochen Cui, Samuel Hwang, Alyssa Pradhan, Kevin Yuan, and David W. Eyre. 2026. BDI at MEDIQA-EVAL 2026: A ReAct-Style Multimodal Agent for Fine-Grained Medical Response Assessment. In Proceedings of the 8th Workshop on Clinical Natural Language Processing (Clinical NLP) @ LREC 2026, pages 141–152, Palma, Mallorca (Spain). ELRA Language Resources Association (ELRA).