@inproceedings{kunle-john-etal-2026-pubmedcausal,
title = "{P}ub{M}ed{C}ausal: A Span-Level Annotated Corpus for Causal Relation Extraction in Biomedical Text",
author = "Kunle-John, Ifeoluwa Oluwatosin and
Josiah, Paul and
Agbaakin, Oluwatosin and
Aina, Peter Olusegun and
Odezuligbo, Ikenna and
Anuyah, Sydney",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1798/",
pages = "40966--40997",
ISBN = "979-8-89176-459-0",
abstract = "Causal relation extraction (CRE) is central to biomedical text mining, but current resources often conflate causal relations with broader associations, restrict annotation to sentence-level examples, or focus mainly on explicit causal cues. This limits their usefulness for evaluating whether models can recover causal claims as they are actually expressed in biomedical text. We introduce PubMedCausal, a span-level annotated corpus for biomedical CRE built from PubMed abstracts. The corpus contains 30,000 paragraph-level rows, including 3,945 causal rows and 6,491 adjudicated cause{--}effect pairs. Each causal relation is annotated with full-text cause and effect spans, causality type, and sententiality, enabling evaluation of both causal detection and full-span causal extraction. We benchmark discriminative encoders and open-source generative models across detection and extraction settings. For causal detection, biomedical encoders are strongest, with PubMedBERT reaching an F$_1$ score of 0.7391. For span-level extraction, the best generative baseline is DeepSeek-R1-32B with few-shot prompting, reaching a Cosine Pair F$_1$ of 0.6765. We further test transfer learning by evaluating PubMedCausal-trained encoders on external causal relation datasets, showing that the resource supports cross-dataset evaluation beyond its original biomedical setting. Our results show that biomedical CRE remains difficult under class imbalance, long causal spans, implicit causality, inter-sentential relations, and prompt sensitivity. PubMedCausal provides a high-quality benchmark for studying causal detection, full-span extraction, and the limits of current language models in biomedical causal reasoning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kunle-john-etal-2026-pubmedcausal">
<titleInfo>
<title>PubMedCausal: A Span-Level Annotated Corpus for Causal Relation Extraction in Biomedical Text</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ifeoluwa</namePart>
<namePart type="given">Oluwatosin</namePart>
<namePart type="family">Kunle-John</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Josiah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Oluwatosin</namePart>
<namePart type="family">Agbaakin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="given">Olusegun</namePart>
<namePart type="family">Aina</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ikenna</namePart>
<namePart type="family">Odezuligbo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sydney</namePart>
<namePart type="family">Anuyah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Causal relation extraction (CRE) is central to biomedical text mining, but current resources often conflate causal relations with broader associations, restrict annotation to sentence-level examples, or focus mainly on explicit causal cues. This limits their usefulness for evaluating whether models can recover causal claims as they are actually expressed in biomedical text. We introduce PubMedCausal, a span-level annotated corpus for biomedical CRE built from PubMed abstracts. The corpus contains 30,000 paragraph-level rows, including 3,945 causal rows and 6,491 adjudicated cause–effect pairs. Each causal relation is annotated with full-text cause and effect spans, causality type, and sententiality, enabling evaluation of both causal detection and full-span causal extraction. We benchmark discriminative encoders and open-source generative models across detection and extraction settings. For causal detection, biomedical encoders are strongest, with PubMedBERT reaching an F₁ score of 0.7391. For span-level extraction, the best generative baseline is DeepSeek-R1-32B with few-shot prompting, reaching a Cosine Pair F₁ of 0.6765. We further test transfer learning by evaluating PubMedCausal-trained encoders on external causal relation datasets, showing that the resource supports cross-dataset evaluation beyond its original biomedical setting. Our results show that biomedical CRE remains difficult under class imbalance, long causal spans, implicit causality, inter-sentential relations, and prompt sensitivity. PubMedCausal provides a high-quality benchmark for studying causal detection, full-span extraction, and the limits of current language models in biomedical causal reasoning.</abstract>
<identifier type="citekey">kunle-john-etal-2026-pubmedcausal</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1798/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>40966</start>
<end>40997</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PubMedCausal: A Span-Level Annotated Corpus for Causal Relation Extraction in Biomedical Text
%A Kunle-John, Ifeoluwa Oluwatosin
%A Josiah, Paul
%A Agbaakin, Oluwatosin
%A Aina, Peter Olusegun
%A Odezuligbo, Ikenna
%A Anuyah, Sydney
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F kunle-john-etal-2026-pubmedcausal
%X Causal relation extraction (CRE) is central to biomedical text mining, but current resources often conflate causal relations with broader associations, restrict annotation to sentence-level examples, or focus mainly on explicit causal cues. This limits their usefulness for evaluating whether models can recover causal claims as they are actually expressed in biomedical text. We introduce PubMedCausal, a span-level annotated corpus for biomedical CRE built from PubMed abstracts. The corpus contains 30,000 paragraph-level rows, including 3,945 causal rows and 6,491 adjudicated cause–effect pairs. Each causal relation is annotated with full-text cause and effect spans, causality type, and sententiality, enabling evaluation of both causal detection and full-span causal extraction. We benchmark discriminative encoders and open-source generative models across detection and extraction settings. For causal detection, biomedical encoders are strongest, with PubMedBERT reaching an F₁ score of 0.7391. For span-level extraction, the best generative baseline is DeepSeek-R1-32B with few-shot prompting, reaching a Cosine Pair F₁ of 0.6765. We further test transfer learning by evaluating PubMedCausal-trained encoders on external causal relation datasets, showing that the resource supports cross-dataset evaluation beyond its original biomedical setting. Our results show that biomedical CRE remains difficult under class imbalance, long causal spans, implicit causality, inter-sentential relations, and prompt sensitivity. PubMedCausal provides a high-quality benchmark for studying causal detection, full-span extraction, and the limits of current language models in biomedical causal reasoning.
%U https://aclanthology.org/2026.emnlp-main.1798/
%P 40966-40997
Markdown (Informal)
[PubMedCausal: A Span-Level Annotated Corpus for Causal Relation Extraction in Biomedical Text](https://aclanthology.org/2026.emnlp-main.1798/) (Kunle-John et al., EMNLP 2026)
ACL