@inproceedings{shimizu-etal-2026-herd,
title = "A Herd of Language Models Makes a Better Zero-shot Annotator for Clinical Named Entity Recognition",
author = "Shimizu, Seiji and
Wakamiya, Shoko and
Aramaki, Eiji",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.599/",
pages = "12327--12344",
ISBN = "979-8-89176-395-1",
abstract = "Clinical named entity recognition (NER) remains difficult to scale due to the high cost of manual annotation. Although large language models (LLMs) enable zero-shot annotation, their performance on clinical NER is still limited. To this end, we improve the annotation quality by aggregating annotations from *a herd of diverse LLMs*, including general-purpose, medically adapted, and NER-specialized models. A key challenge in this multi-LLM setting is effectively leveraging entities extracted by only a minority of models: although they account for a substantial portion of true positives, they are heavily intermixed with noise. To address this, we introduce **MARY**, a label-modeling method for **M**ulti-LLM **A**nnotation using **R**epresentation learning to capture contextual similarit**Y**. During aggregation, MARY selectively incorporates minority-extracted entities whose contexts are similar to those of majority-extracted entities, yielding more reliable and comprehensive annotations. Experimental results show that MARY improves the average F1 score by 8.6{\%} over vanilla zero-shot baselines while reducing annotation costs."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shimizu-etal-2026-herd">
<titleInfo>
<title>A Herd of Language Models Makes a Better Zero-shot Annotator for Clinical Named Entity Recognition</title>
</titleInfo>
<name type="personal">
<namePart type="given">Seiji</namePart>
<namePart type="family">Shimizu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shoko</namePart>
<namePart type="family">Wakamiya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eiji</namePart>
<namePart type="family">Aramaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Clinical named entity recognition (NER) remains difficult to scale due to the high cost of manual annotation. Although large language models (LLMs) enable zero-shot annotation, their performance on clinical NER is still limited. To this end, we improve the annotation quality by aggregating annotations from *a herd of diverse LLMs*, including general-purpose, medically adapted, and NER-specialized models. A key challenge in this multi-LLM setting is effectively leveraging entities extracted by only a minority of models: although they account for a substantial portion of true positives, they are heavily intermixed with noise. To address this, we introduce **MARY**, a label-modeling method for **M**ulti-LLM **A**nnotation using **R**epresentation learning to capture contextual similarit**Y**. During aggregation, MARY selectively incorporates minority-extracted entities whose contexts are similar to those of majority-extracted entities, yielding more reliable and comprehensive annotations. Experimental results show that MARY improves the average F1 score by 8.6% over vanilla zero-shot baselines while reducing annotation costs.</abstract>
<identifier type="citekey">shimizu-etal-2026-herd</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.599/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>12327</start>
<end>12344</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Herd of Language Models Makes a Better Zero-shot Annotator for Clinical Named Entity Recognition
%A Shimizu, Seiji
%A Wakamiya, Shoko
%A Aramaki, Eiji
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F shimizu-etal-2026-herd
%X Clinical named entity recognition (NER) remains difficult to scale due to the high cost of manual annotation. Although large language models (LLMs) enable zero-shot annotation, their performance on clinical NER is still limited. To this end, we improve the annotation quality by aggregating annotations from *a herd of diverse LLMs*, including general-purpose, medically adapted, and NER-specialized models. A key challenge in this multi-LLM setting is effectively leveraging entities extracted by only a minority of models: although they account for a substantial portion of true positives, they are heavily intermixed with noise. To address this, we introduce **MARY**, a label-modeling method for **M**ulti-LLM **A**nnotation using **R**epresentation learning to capture contextual similarit**Y**. During aggregation, MARY selectively incorporates minority-extracted entities whose contexts are similar to those of majority-extracted entities, yielding more reliable and comprehensive annotations. Experimental results show that MARY improves the average F1 score by 8.6% over vanilla zero-shot baselines while reducing annotation costs.
%U https://aclanthology.org/2026.findings-acl.599/
%P 12327-12344
Markdown (Informal)
[A Herd of Language Models Makes a Better Zero-shot Annotator for Clinical Named Entity Recognition](https://aclanthology.org/2026.findings-acl.599/) (Shimizu et al., Findings 2026)
ACL