<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="danu-2026-siemens">
    <titleInfo>
        <title>SIEMENS at #SMM4H–HeaRD 2026: The Impact of Training Strategy and Backbone Selection on BERT-based Multilingual Clinical NER</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Manuela</namePart>
        <namePart type="given">Daniela</namePart>
        <namePart type="family">Danu</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>Proceedings of the 11th Social Media Mining for Health Research and Applications (SMM4H-HeaRD 2026) Workshop and Shared Tasks</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Guillermo</namePart>
            <namePart type="family">Lopez-Garcia</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Graciela</namePart>
            <namePart type="family">Gonzalez-Hernandez</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, United States</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-432-3</identifier>
    </relatedItem>
    <abstract>This paper describes our participation in the MultiClinNER subtask of the MultiClinAI shared task, part of the #SMM4H-HeaRD Workshop at ACL 2026. The task requires identifying DISEASE, SYMPTOM, and PROCEDURE mentions in clinical case reports across seven languages: Czech, Dutch, English, Italian, Romanian, Spanish, and Swedish. We compare two BERT-based sequence labeling methods: (i) sentence-level token classification with a fixed train/validation split, and (ii) paragraph-level chunking with 5-fold cross-validation and checkpoint merging, using language-specific BERT models and multilingual XLM-RoBERTa-large as backbones. Our results show that 5-fold training with checkpoint merging consistently outperforms the fixed split strategy, with further analysis suggesting that the gains are primarily driven by improved training-set coverage rather than by differences in input granularity. Language-specific BERT encoders prove most effective for Spanish and English, while XLM-RoBERTa-large yields the strongest results for the remaining five languages through cross-lingual transfer.</abstract>
    <identifier type="citekey">danu-2026-siemens</identifier>
    <location>
        <url>https://aclanthology.org/2026.smm4h-1.34/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>216</start>
            <end>221</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T SIEMENS at #SMM4H–HeaRD 2026: The Impact of Training Strategy and Backbone Selection on BERT-based Multilingual Clinical NER
%A Danu, Manuela Daniela
%Y Lopez-Garcia, Guillermo
%Y Gonzalez-Hernandez, Graciela
%S Proceedings of the 11th Social Media Mining for Health Research and Applications (SMM4H-HeaRD 2026) Workshop and Shared Tasks
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, United States
%@ 979-8-89176-432-3
%F danu-2026-siemens
%X This paper describes our participation in the MultiClinNER subtask of the MultiClinAI shared task, part of the #SMM4H-HeaRD Workshop at ACL 2026. The task requires identifying DISEASE, SYMPTOM, and PROCEDURE mentions in clinical case reports across seven languages: Czech, Dutch, English, Italian, Romanian, Spanish, and Swedish. We compare two BERT-based sequence labeling methods: (i) sentence-level token classification with a fixed train/validation split, and (ii) paragraph-level chunking with 5-fold cross-validation and checkpoint merging, using language-specific BERT models and multilingual XLM-RoBERTa-large as backbones. Our results show that 5-fold training with checkpoint merging consistently outperforms the fixed split strategy, with further analysis suggesting that the gains are primarily driven by improved training-set coverage rather than by differences in input granularity. Language-specific BERT encoders prove most effective for Spanish and English, while XLM-RoBERTa-large yields the strongest results for the remaining five languages through cross-lingual transfer.
%U https://aclanthology.org/2026.smm4h-1.34/
%P 216-221

Markdown (Informal)

[SIEMENS at #SMM4H–HeaRD 2026: The Impact of Training Strategy and Backbone Selection on BERT-based Multilingual Clinical NER](https://aclanthology.org/2026.smm4h-1.34/) (Danu, SMM4H 2026)

• SIEMENS at #SMM4H–HeaRD 2026: The Impact of Training Strategy and Backbone Selection on BERT-based Multilingual Clinical NER (Danu, SMM4H 2026)

ACL

• Manuela Daniela Danu. 2026. SIEMENS at #SMM4H–HeaRD 2026: The Impact of Training Strategy and Backbone Selection on BERT-based Multilingual Clinical NER. In Proceedings of the 11th Social Media Mining for Health Research and Applications (SMM4H-HeaRD 2026) Workshop and Shared Tasks, pages 216–221, San Diego, United States. Association for Computational Linguistics.