@inproceedings{danoe-etal-2026-dr,
title = "Dr-{BERT}-{NL} at {\#}{SMM}4{H}{--}{H}ea{RD} 2026: {DOKTERBERT} {--} Ontology-Grounded Contextual Representations for {D}utch Clinical {NLP}",
author = "Danoe, Gijs and
Voss, Andreas and
Hamprecht, Axel and
Berends, Matthijs S.",
editor = "Lopez-Garcia, Guillermo and
Gonzalez-Hernandez, Graciela",
booktitle = "Proceedings of the 11th Social Media Mining for Health Research and Applications ({SMM}4{H}-{H}ea{RD} 2026) Workshop and Shared Tasks",
month = jul,
year = "2026",
address = "San Diego, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.smm4h-1.25/",
pages = "154--159",
ISBN = "979-8-89176-432-3",
abstract = "We describe our submission to SMM4H-HeaRD 2026 Task 7, which asks systems tolabel ClinicalImpacts and SocialImpactsspans in Reddit posts about non-medical sub-stance use. We compare four pipeline shapesbuilt on the same DeBERTa-v3-base back-bone: (i) a direct 5-class encoder with a linear-chain CRF head, (ii) a two-stage detect-then-classify pipeline that delegates span typingto an instruction-tuned LLM (Qwen2.5-7Bor Gemma-3-12B, 4-bit NF4), (iii) an auditpipeline in which the same LLM verifies theencoder{'}s predictions, and (iv) a classical-MLvariant that replaces the LLM with an SVMtrained on encoder span embeddings. Across16 configurations, the encoder-only DeBERTa-v3 + CRF configuration is the strongest sin-gle system on the official test split, reaching45.4{\%} strict and 54.2{\%} relaxed F1 {---} +8.6/ +5.3 points above a mental-roberta-basebaseline. LLM audits give a small dev gain thatdoes not transfer to test."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="danoe-etal-2026-dr">
<titleInfo>
<title>Dr-BERT-NL at #SMM4H–HeaRD 2026: DOKTERBERT – Ontology-Grounded Contextual Representations for Dutch Clinical NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gijs</namePart>
<namePart type="family">Danoe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andreas</namePart>
<namePart type="family">Voss</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Axel</namePart>
<namePart type="family">Hamprecht</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Matthijs</namePart>
<namePart type="given">S</namePart>
<namePart type="family">Berends</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>We describe our submission to SMM4H-HeaRD 2026 Task 7, which asks systems tolabel ClinicalImpacts and SocialImpactsspans in Reddit posts about non-medical sub-stance use. We compare four pipeline shapesbuilt on the same DeBERTa-v3-base back-bone: (i) a direct 5-class encoder with a linear-chain CRF head, (ii) a two-stage detect-then-classify pipeline that delegates span typingto an instruction-tuned LLM (Qwen2.5-7Bor Gemma-3-12B, 4-bit NF4), (iii) an auditpipeline in which the same LLM verifies theencoder’s predictions, and (iv) a classical-MLvariant that replaces the LLM with an SVMtrained on encoder span embeddings. Across16 configurations, the encoder-only DeBERTa-v3 + CRF configuration is the strongest sin-gle system on the official test split, reaching45.4% strict and 54.2% relaxed F1 — +8.6/ +5.3 points above a mental-roberta-basebaseline. LLM audits give a small dev gain thatdoes not transfer to test.</abstract>
<identifier type="citekey">danoe-etal-2026-dr</identifier>
<location>
<url>https://aclanthology.org/2026.smm4h-1.25/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>154</start>
<end>159</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Dr-BERT-NL at #SMM4H–HeaRD 2026: DOKTERBERT – Ontology-Grounded Contextual Representations for Dutch Clinical NLP
%A Danoe, Gijs
%A Voss, Andreas
%A Hamprecht, Axel
%A Berends, Matthijs S.
%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 danoe-etal-2026-dr
%X We describe our submission to SMM4H-HeaRD 2026 Task 7, which asks systems tolabel ClinicalImpacts and SocialImpactsspans in Reddit posts about non-medical sub-stance use. We compare four pipeline shapesbuilt on the same DeBERTa-v3-base back-bone: (i) a direct 5-class encoder with a linear-chain CRF head, (ii) a two-stage detect-then-classify pipeline that delegates span typingto an instruction-tuned LLM (Qwen2.5-7Bor Gemma-3-12B, 4-bit NF4), (iii) an auditpipeline in which the same LLM verifies theencoder’s predictions, and (iv) a classical-MLvariant that replaces the LLM with an SVMtrained on encoder span embeddings. Across16 configurations, the encoder-only DeBERTa-v3 + CRF configuration is the strongest sin-gle system on the official test split, reaching45.4% strict and 54.2% relaxed F1 — +8.6/ +5.3 points above a mental-roberta-basebaseline. LLM audits give a small dev gain thatdoes not transfer to test.
%U https://aclanthology.org/2026.smm4h-1.25/
%P 154-159
Markdown (Informal)
[Dr-BERT-NL at #SMM4H–HeaRD 2026: DOKTERBERT – Ontology-Grounded Contextual Representations for Dutch Clinical NLP](https://aclanthology.org/2026.smm4h-1.25/) (Danoe et al., SMM4H 2026)
ACL