EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Integrating deep learning architectures for enhanced biomedical relation extraction: a pipeline approach

M Janina Sarol iD, Gibong Hong, Evan Guerra, Halil Kilicoglu iD

Database · 2024

DOI: 10.1093/database/baae079 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Database, 2024. Verfasst von M Janina Sarol und 3 weiteren Autor:innen. Beteiligte Einrichtung: Informatics Programs, University of Illinois Urbana-Champaign, 614 E Daniel Street, Champaign, IL 61820, United States. Hauptthemen laut MEDLINE: Deep Learning, Natural Language Processing. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 66 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NLM NIH HHS, ORI HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Biomedical relation extraction from scientific publications is a key task in biomedical natural language processing (NLP) and can facilitate the creation of large knowledge bases, enable more efficient knowledge discovery, and accelerate evidence synthesis. In this paper, building upon our previous effort in the BioCreative VIII BioRED Track, we propose an enhanced end-to-end pipeline approach for biomedical relation extraction (RE) and novelty detection (ND) that effectively leverages existing datasets and integrates state-of-the-art deep learning methods. Our pipeline consists of four tasks performed sequentially: named entity recognition (NER), entity linking (EL), RE, and ND. We trained models using the BioRED benchmark corpus that was the basis of the shared task. We explored several methods for each task and combinations thereof: for NER, we compared a BERT-based sequence labeling model that uses the BIO scheme with a span classification model. For EL, we trained a convolutional neural network model for diseases and chemicals and used an existing tool, PubTator 3.0, for mapping other entity types. For RE and ND, we adapted the BERT-based, sentence-bound PURE model to bidirectional and document-level extraction. We also performed extensive hyperparameter tuning to improve model performance. We obtained our best performance using BERT-based models for NER, RE, and ND, and the hybrid approach for EL. Our enhanced and optimized pipeline showed substantial improvement compared to our shared task submission, NER: 93.53 (+3.09), EL: 83.87 (+9.73), RE: 46.18 (+15.67), and ND: 38.86 (+14.9). While the performances of the NER and EL models are reasonably high, RE and ND tasks remain challenging at the document level. Further enhancements to the dataset could enable more accurate and useful models for practical use. We provide our models and code at https://github.com/janinaj/e2eBioMedRE/. Database URL: https://github.com/janinaj/e2eBioMedRE/

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningNatural Language ProcessingData MiningNeural Networks, ComputerDatabases, Factual

Autor:innen und Institutionen

  • Informatics Programs, University of Illinois Urbana-Champaign, 614 E Daniel Street, Champaign, IL 61820, United States.
  • School of Information Sciences, University of Illinois Urbana-Champaign, 501 E Daniel Street, Champaign, IL 61820, United States.

ORCID-Profile: M Janina Sarol, Halil Kilicoglu

Förderung

  • NLM NIH HHS
  • ORI HHS

Zitieren

APA

Sarol, M. J., Hong, G., Guerra, E., & Kilicoglu, H. (2024). Integrating deep learning architectures for enhanced biomedical relation extraction: a pipeline approach. Database. https://doi.org/10.1093/database/baae079

Vancouver

Sarol MJ, Hong G, Guerra E, Kilicoglu H. Integrating deep learning architectures for enhanced biomedical relation extraction: a pipeline approach. Database. 2024. doi:10.1093/database/baae079

BibTeX
@article{Sarol2024,
  author = {Sarol, M Janina and Hong, Gibong and Guerra, Evan and Kilicoglu, Halil},
  title = {Integrating deep learning architectures for enhanced biomedical relation extraction: a pipeline approach},
  journal = {Database},
  year = {2024},
  doi = {10.1093/database/baae079}
}
Kostenlos starten →