TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications

Sepideh Mesbah; Christoph Lofi; Manuel Valle Torre; Alessandro Bozzon; Geert-Jan Houben

doi:10.1007/978-3-030-00671-6_8

TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications

Sepideh Mesbah, Christoph Lofi, Manuel Valle Torre, Alessandro Bozzon, Geert-Jan Houben

Web Information Systems

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

18 Citations (Scopus)

158 Downloads (Pure)

Abstract

Named Entity Recognition and Typing (NER/NET) is a challenging task, especially with long-tail entities such as the ones found in scientific publications. These entities (e.g. “WebKB”, “StatSnowball”) are rare, often relevant only in specific knowledge domains, yet important for retrieval and exploration purposes. State-of-the-art NER approaches employ supervised machine learning models, trained on expensive typelabeled data laboriously produced by human annotators. A common workaround is the generation of labeled training data from knowledge bases; this approach is not suitable for long-tail entity types that are, by definition, scarcely represented in KBs.
This paper presents an iterative approach for training NER and NET
classifiers in scientific publications that relies on minimal human input,
namely a small seed set of instances for the targeted entity type. We
introduce different strategies for training data extraction, semantic expansion, and result entity filtering.We evaluate our approach on scientific
publications, focusing on the long-tail entities types Datasets, Methods in
computer science publications, and Proteins in biomedical publications.

Original language	English
Title of host publication	The Semantic Web – ISWC 2018
Subtitle of host publication	Proceedings of the 17th International Semantic Web Conference
Editors	D. Vrandečić, K. Bontcheva, M.C. Suárez-Figueroa, V. Presutti, I. Celino, M. Sabou, L.M Kaffee, E. Simperl
Place of Publication	Cham
Publisher	Springer
Pages	127-143
Number of pages	16
ISBN (Electronic)	978-3-030-00671-6
ISBN (Print)	978-3-030-00670-9
DOIs	https://doi.org/10.1007/978-3-030-00671-6_8
Publication status	Published - 2018
Event	ISWC 2018: 17th International Semantic Web Conference - Monterey, CA, United States Duration: 8 Oct 2018 → 12 Oct 2018 Conference number: 17 http://iswc2018.semanticweb.org/

Publication series

Name	Lecture Notes in Computer Science (LNCS)
Volume	11136

Conference

Conference	ISWC 2018
Country/Territory	United States
City	Monterey, CA
Period	8/10/18 → 12/10/18
Internet address	http://iswc2018.semanticweb.org/

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Access to Document

10.1007/978-3-030-00671-6_8

Mesbah2018_Chapter_TSE-NERAnIterativeApproachForLFinal published version, 832 KB

Cite this

Mesbah, S., Lofi, C., Valle Torre, M., Bozzon, A., & Houben, G.-J. (2018). TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications. In D. Vrandečić, K. Bontcheva, M. C. Suárez-Figueroa, V. Presutti, I. Celino, M. Sabou, L. M. Kaffee, & E. Simperl (Eds.), The Semantic Web – ISWC 2018: Proceedings of the 17th International Semantic Web Conference (pp. 127-143). (Lecture Notes in Computer Science (LNCS); Vol. 11136). Springer. https://doi.org/10.1007/978-3-030-00671-6_8

Mesbah, Sepideh ; Lofi, Christoph ; Valle Torre, Manuel et al. / TSE-NER : An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications. The Semantic Web – ISWC 2018: Proceedings of the 17th International Semantic Web Conference. editor / D. Vrandečić ; K. Bontcheva ; M.C. Suárez-Figueroa ; V. Presutti ; I. Celino ; M. Sabou ; L.M Kaffee ; E. Simperl. Cham : Springer, 2018. pp. 127-143 (Lecture Notes in Computer Science (LNCS)).

@inproceedings{91b0bf6013044b2fba55f58f04351381,

title = "TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications",

abstract = "Named Entity Recognition and Typing (NER/NET) is a challenging task, especially with long-tail entities such as the ones found in scientific publications. These entities (e.g. “WebKB”, “StatSnowball”) are rare, often relevant only in specific knowledge domains, yet important for retrieval and exploration purposes. State-of-the-art NER approaches employ supervised machine learning models, trained on expensive typelabeled data laboriously produced by human annotators. A common workaround is the generation of labeled training data from knowledge bases; this approach is not suitable for long-tail entity types that are, by definition, scarcely represented in KBs.This paper presents an iterative approach for training NER and NETclassifiers in scientific publications that relies on minimal human input,namely a small seed set of instances for the targeted entity type. Weintroduce different strategies for training data extraction, semantic expansion, and result entity filtering.We evaluate our approach on scientificpublications, focusing on the long-tail entities types Datasets, Methods incomputer science publications, and Proteins in biomedical publications.",

author = "Sepideh Mesbah and Christoph Lofi and {Valle Torre}, Manuel and Alessandro Bozzon and Geert-Jan Houben",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.; ISWC 2018 : 17th International Semantic Web Conference ; Conference date: 08-10-2018 Through 12-10-2018",

year = "2018",

doi = "10.1007/978-3-030-00671-6_8",

language = "English",

isbn = "978-3-030-00670-9",

series = "Lecture Notes in Computer Science (LNCS)",

publisher = "Springer",

pages = "127--143",

editor = "{ Vrande{\v c}i{\'c}}, D. and K. Bontcheva and M.C. Su{\'a}rez-Figueroa and V. Presutti and I. Celino and M. Sabou and L.M Kaffee and E. Simperl",

booktitle = "The Semantic Web – ISWC 2018",

url = "http://iswc2018.semanticweb.org/",

}

Mesbah, S, Lofi, C , Valle Torre, M , Bozzon, A & Houben, G-J 2018, TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications. in D Vrandečić, K Bontcheva, MC Suárez-Figueroa, V Presutti, I Celino, M Sabou, LM Kaffee & E Simperl (eds), The Semantic Web – ISWC 2018: Proceedings of the 17th International Semantic Web Conference. Lecture Notes in Computer Science (LNCS), vol. 11136, Springer, Cham, pp. 127-143, ISWC 2018, Monterey, CA, United States, 8/10/18. https://doi.org/10.1007/978-3-030-00671-6_8

TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications. / Mesbah, Sepideh; Lofi, Christoph ; Valle Torre, Manuel et al.
The Semantic Web – ISWC 2018: Proceedings of the 17th International Semantic Web Conference. ed. / D. Vrandečić; K. Bontcheva; M.C. Suárez-Figueroa; V. Presutti; I. Celino; M. Sabou; L.M Kaffee; E. Simperl. Cham: Springer, 2018. p. 127-143 (Lecture Notes in Computer Science (LNCS); Vol. 11136).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - TSE-NER

T2 - ISWC 2018

AU - Mesbah, Sepideh

AU - Lofi, Christoph

AU - Valle Torre, Manuel

AU - Bozzon, Alessandro

AU - Houben, Geert-Jan

N1 - Conference code: 17

PY - 2018

Y1 - 2018

N2 - Named Entity Recognition and Typing (NER/NET) is a challenging task, especially with long-tail entities such as the ones found in scientific publications. These entities (e.g. “WebKB”, “StatSnowball”) are rare, often relevant only in specific knowledge domains, yet important for retrieval and exploration purposes. State-of-the-art NER approaches employ supervised machine learning models, trained on expensive typelabeled data laboriously produced by human annotators. A common workaround is the generation of labeled training data from knowledge bases; this approach is not suitable for long-tail entity types that are, by definition, scarcely represented in KBs.This paper presents an iterative approach for training NER and NETclassifiers in scientific publications that relies on minimal human input,namely a small seed set of instances for the targeted entity type. Weintroduce different strategies for training data extraction, semantic expansion, and result entity filtering.We evaluate our approach on scientificpublications, focusing on the long-tail entities types Datasets, Methods incomputer science publications, and Proteins in biomedical publications.

AB - Named Entity Recognition and Typing (NER/NET) is a challenging task, especially with long-tail entities such as the ones found in scientific publications. These entities (e.g. “WebKB”, “StatSnowball”) are rare, often relevant only in specific knowledge domains, yet important for retrieval and exploration purposes. State-of-the-art NER approaches employ supervised machine learning models, trained on expensive typelabeled data laboriously produced by human annotators. A common workaround is the generation of labeled training data from knowledge bases; this approach is not suitable for long-tail entity types that are, by definition, scarcely represented in KBs.This paper presents an iterative approach for training NER and NETclassifiers in scientific publications that relies on minimal human input,namely a small seed set of instances for the targeted entity type. Weintroduce different strategies for training data extraction, semantic expansion, and result entity filtering.We evaluate our approach on scientificpublications, focusing on the long-tail entities types Datasets, Methods incomputer science publications, and Proteins in biomedical publications.

U2 - 10.1007/978-3-030-00671-6_8

DO - 10.1007/978-3-030-00671-6_8

M3 - Conference contribution

SN - 978-3-030-00670-9

T3 - Lecture Notes in Computer Science (LNCS)

SP - 127

EP - 143

BT - The Semantic Web – ISWC 2018

A2 - Vrandečić, D.

A2 - Bontcheva, K.

A2 - Suárez-Figueroa, M.C.

A2 - Presutti, V.

A2 - Celino, I.

A2 - Sabou, M.

A2 - Kaffee, L.M

A2 - Simperl, E.

PB - Springer

CY - Cham

Y2 - 8 October 2018 through 12 October 2018

ER -

Mesbah S, Lofi C , Valle Torre M , Bozzon A , Houben GJ. TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications. In Vrandečić D, Bontcheva K, Suárez-Figueroa MC, Presutti V, Celino I, Sabou M, Kaffee LM, Simperl E, editors, The Semantic Web – ISWC 2018: Proceedings of the 17th International Semantic Web Conference. Cham: Springer. 2018. p. 127-143. (Lecture Notes in Computer Science (LNCS)). doi: 10.1007/978-3-030-00671-6_8

TSE-NER: An Iterative Approach for Long-Tail Entity Extraction in Scientific Publications

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Fingerprint

Cite this