Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings

In this paper, we explore the application of artificial neural network (‘deep learning’) methods to the problem of detecting chemical-protein interactions in PubMed abstracts. We present here a system using multiple Long Short Term Memory layers to analyse candidate interactions, to determine whethe...

Descripción completa

Detalles Bibliográficos
Autores principales: Corbett, P, Boyle, J
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Oxford University Press 2018
Materias:
Original Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6044291/
https://www.ncbi.nlm.nih.gov/pubmed/30010749
http://dx.doi.org/10.1093/database/bay066
_version_ 1783339455599345664
author Corbett, P
Boyle, J
author_facet Corbett, P
Boyle, J
author_sort Corbett, P
collection PubMed
description In this paper, we explore the application of artificial neural network (‘deep learning’) methods to the problem of detecting chemical-protein interactions in PubMed abstracts. We present here a system using multiple Long Short Term Memory layers to analyse candidate interactions, to determine whether there is a relation and which type. A particular feature of our system is the use of unlabelled data, both to pre-train word embeddings and also pre-train LSTM layers in the neural network. On the BioCreative VI CHEMPROT test corpus, our system achieves an F score of 61.51% (56.10% precision, 67.84% recall).
format Online
Article
Text
id pubmed-6044291
institution National Center for Biotechnology Information
language English
publishDate 2018
publisher Oxford University Press
record_format MEDLINE/PubMed
spelling pubmed-60442912018-07-19 Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings Corbett, P Boyle, J Database (Oxford) Original Article In this paper, we explore the application of artificial neural network (‘deep learning’) methods to the problem of detecting chemical-protein interactions in PubMed abstracts. We present here a system using multiple Long Short Term Memory layers to analyse candidate interactions, to determine whether there is a relation and which type. A particular feature of our system is the use of unlabelled data, both to pre-train word embeddings and also pre-train LSTM layers in the neural network. On the BioCreative VI CHEMPROT test corpus, our system achieves an F score of 61.51% (56.10% precision, 67.84% recall). Oxford University Press 2018-07-12 /pmc/articles/PMC6044291/ /pubmed/30010749 http://dx.doi.org/10.1093/database/bay066 Text en © The Author(s) 2018. Published by Oxford University Press. http://creativecommons.org/licenses/by/4.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
spellingShingle Original Article
Corbett, P
Boyle, J
Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title_full Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title_fullStr Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title_full_unstemmed Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title_short Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
title_sort improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings
topic Original Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6044291/
https://www.ncbi.nlm.nih.gov/pubmed/30010749
http://dx.doi.org/10.1093/database/bay066
work_keys_str_mv AT corbettp improvingthelearningofchemicalproteininteractionsfromliteratureusingtransferlearningandspecializedwordembeddings
AT boylej improvingthelearningofchemicalproteininteractionsfromliteratureusingtransferlearningandspecializedwordembeddings

Ejemplares similares