@inproceedings{torri-ieva-2026-polimi,
title = "Polimi at {CRF} Filling 2026: Prompt-Based Information Extraction from {I}talian Clinical Notes",
author = "Torri, Vittorio and
Ieva, Francesca",
editor = "Gupta, Deepak and
Thompson, Paul and
Ananiadou, Sophia and
Demner-Fushman, Dina",
booktitle = "Proceedings of the Third Workshop on Patient-Oriented Language Processing ({CL}4{H}ealth) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.cl4health-1.38/",
doi = "10.63317/5moj6trbujsb",
pages = "412--427",
abstract = "In this paper we describe the system developed by the Polimi team for the CRF Filling Shared Task 2026, which focuses on extracting structured variables from clinical notes. The task is challenging due to scarce annotations, heterogeneous clinical language, and the sparsity of the 134 items to be extracted. Our approach relies on prompt-based information extraction using locally deployed open-weight Large Language Models (LLMs). We focused on the Italian subset of the dataset. The pipeline performs zero-shot extraction using task-specific prompts augmented with a glossary of abbreviations derived from unlabeled notes. To improve reliability and reduce hallucinations, the extraction schema is decomposed into multiple prompts targeting groups of variables, whose outputs are merged and refined through deterministic post-processing rules to normalize values and recover missing labels. During development we explored verification stages based on LLM-based prediction validation and synthetic example generation, but these strategies did not improve performance and were not included in the final system. On the development set, the best configuration based on Mistral Small 3.2 24B Instruct achieved an F1-score of 67.51{\%}. On the official test set, our system ranked third overall and second among systems evaluated on the Italian subset, achieving an F1-score of 63{\%}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="torri-ieva-2026-polimi">
<titleInfo>
<title>Polimi at CRF Filling 2026: Prompt-Based Information Extraction from Italian Clinical Notes</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vittorio</namePart>
<namePart type="family">Torri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Francesca</namePart>
<namePart type="family">Ieva</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 Third Workshop on Patient-Oriented Language Processing (CL4Health) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Deepak</namePart>
<namePart type="family">Gupta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Thompson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sophia</namePart>
<namePart type="family">Ananiadou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dina</namePart>
<namePart type="family">Demner-Fushman</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>In this paper we describe the system developed by the Polimi team for the CRF Filling Shared Task 2026, which focuses on extracting structured variables from clinical notes. The task is challenging due to scarce annotations, heterogeneous clinical language, and the sparsity of the 134 items to be extracted. Our approach relies on prompt-based information extraction using locally deployed open-weight Large Language Models (LLMs). We focused on the Italian subset of the dataset. The pipeline performs zero-shot extraction using task-specific prompts augmented with a glossary of abbreviations derived from unlabeled notes. To improve reliability and reduce hallucinations, the extraction schema is decomposed into multiple prompts targeting groups of variables, whose outputs are merged and refined through deterministic post-processing rules to normalize values and recover missing labels. During development we explored verification stages based on LLM-based prediction validation and synthetic example generation, but these strategies did not improve performance and were not included in the final system. On the development set, the best configuration based on Mistral Small 3.2 24B Instruct achieved an F1-score of 67.51%. On the official test set, our system ranked third overall and second among systems evaluated on the Italian subset, achieving an F1-score of 63%.</abstract>
<identifier type="citekey">torri-ieva-2026-polimi</identifier>
<identifier type="doi">10.63317/5moj6trbujsb</identifier>
<location>
<url>https://aclanthology.org/2026.cl4health-1.38/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>412</start>
<end>427</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Polimi at CRF Filling 2026: Prompt-Based Information Extraction from Italian Clinical Notes
%A Torri, Vittorio
%A Ieva, Francesca
%Y Gupta, Deepak
%Y Thompson, Paul
%Y Ananiadou, Sophia
%Y Demner-Fushman, Dina
%S Proceedings of the Third Workshop on Patient-Oriented Language Processing (CL4Health) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F torri-ieva-2026-polimi
%X In this paper we describe the system developed by the Polimi team for the CRF Filling Shared Task 2026, which focuses on extracting structured variables from clinical notes. The task is challenging due to scarce annotations, heterogeneous clinical language, and the sparsity of the 134 items to be extracted. Our approach relies on prompt-based information extraction using locally deployed open-weight Large Language Models (LLMs). We focused on the Italian subset of the dataset. The pipeline performs zero-shot extraction using task-specific prompts augmented with a glossary of abbreviations derived from unlabeled notes. To improve reliability and reduce hallucinations, the extraction schema is decomposed into multiple prompts targeting groups of variables, whose outputs are merged and refined through deterministic post-processing rules to normalize values and recover missing labels. During development we explored verification stages based on LLM-based prediction validation and synthetic example generation, but these strategies did not improve performance and were not included in the final system. On the development set, the best configuration based on Mistral Small 3.2 24B Instruct achieved an F1-score of 67.51%. On the official test set, our system ranked third overall and second among systems evaluated on the Italian subset, achieving an F1-score of 63%.
%R 10.63317/5moj6trbujsb
%U https://aclanthology.org/2026.cl4health-1.38/
%U https://doi.org/10.63317/5moj6trbujsb
%P 412-427
Markdown (Informal)
[Polimi at CRF Filling 2026: Prompt-Based Information Extraction from Italian Clinical Notes](https://aclanthology.org/2026.cl4health-1.38/) (Torri & Ieva, CL4Health 2026)
ACL