MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach

Diego Moussallem; Ricardo Usbeck; Michael Röder; Axel Cyrille Ngonga Ngomo

doi:10.1145/3148011.3148024

MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach

Diego Moussallem^*
, Ricardo Usbeck
, Michael Röder
, Axel Cyrille Ngonga Ngomo

^*Korrespondierende/r Autor/-in für diese Arbeit

Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › Begutachtung

38 Zitate (Scopus)

Abstract

Entity linking has recently been the subject of a significant body of research. Currently, the best performing approaches rely on trained mono-lingual models. Porting these approaches to other languages is consequently a difficult endeavor as it requires corresponding training data and retraining of the models. We address this drawback by presenting a novel multilingual, knowledge-base agnostic and deterministic approach to entity linking, dubbed MAG. MAG is based on a combination of context-based retrieval on structured knowledge bases and graph algorithms. We evaluate MAG on 23 data sets and in 7 languages. Our results showthat the best approach trained on English datasets (PBOH) achieves a micro F-measure that is up to 4 times worse on datasets in other languages. MAG on the other hand achieves state-of-the-art performance on English datasets and reaches a micro F-measure that is up to 0.6 higher than that of PBOH on non-English languages.

Originalsprache	Englisch
Titel	Proceedings of the Knowledge Capture Conference, K-CAP 2017
Seitenumfang	8
Herausgeber (Verlag)	Association for Computing Machinery, Inc
Erscheinungsdatum	04.12.2017
Seiten	1-8
Aufsatznummer	9
ISBN (Print)	978-1-4503-5553-7
DOIs	https://doi.org/10.1145/3148011.3148024 https://doi.org/10.48550/arXiv.1707.05288
Publikationsstatus	Erschienen - 04.12.2017
Extern publiziert	Ja
Veranstaltung	The 9th International Conference on Knowledge Capture - K-CAP 2017: 9th International Conference on Knowledge Capture - Austin, Texas, USA, Austin, USA / Vereinigte Staaten Dauer: 04.12.2017 → 06.12.2017 Konferenznummer: 9 http://www.k-cap2017.org (Offizielle Event-Webseite) https://k-cap2017.org

Fachgebiete und Schlagwörter

Informatik
Wirtschaftsinformatik

Zugriff auf Dokument

10.1145/3148011.3148024
10.48550/arXiv.1707.05288Lizenz: CC0

Dieses zitieren

Moussallem, D., Usbeck, R., Röder, M., & Ngomo, A. C. N. (2017). MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach. in Proceedings of the Knowledge Capture Conference, K-CAP 2017 (S. 1-8). Artikel 9 (Proceedings of the Knowledge Capture Conference, K-CAP 2017). Association for Computing Machinery, Inc. https://doi.org/10.1145/3148011.3148024, https://doi.org/10.48550/arXiv.1707.05288

@inbook{2232305548424a3a99191f31e81abe92,

title = "MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach",

abstract = "Entity linking has recently been the subject of a significant body of research. Currently, the best performing approaches rely on trained mono-lingual models. Porting these approaches to other languages is consequently a difficult endeavor as it requires corresponding training data and retraining of the models. We address this drawback by presenting a novel multilingual, knowledge-base agnostic and deterministic approach to entity linking, dubbed MAG. MAG is based on a combination of context-based retrieval on structured knowledge bases and graph algorithms. We evaluate MAG on 23 data sets and in 7 languages. Our results showthat the best approach trained on English datasets (PBOH) achieves a micro F-measure that is up to 4 times worse on datasets in other languages. MAG on the other hand achieves state-of-the-art performance on English datasets and reaches a micro F-measure that is up to 0.6 higher than that of PBOH on non-English languages.",

keywords = "Entity Linking, Multilingual, Named Entity Disambiguation, Informatics, Business informatics",

author = "Diego Moussallem and Ricardo Usbeck and Michael R{\"o}der and Ngomo, \{Axel Cyrille Ngonga\}",

note = "This work has been supported by the H2020 project HOBBIT (GA no. 688227) as well as the EuroStars projects DIESEL (no. 01QE1512C) and QAMEL (no. 01QE1549C) and supported by the Brazilian National Council for Scientific and Technological Development (CNPq) (no. 206971/2014-1). This work has also been supported by the German Federal Ministry of Transport and Digital Infrastructure (BMVI) in the projects LIMBO (no. 19F2029I) and OPAL (no. 19F2028A) as well as by the German Federal Ministry of Education and Research (BMBF) within {\textquoteright}KMU-innovativ: Forschung f{\"u}r die zivile Sicherheit{\textquoteright} in particular {\textquoteright}Forschung f{\"u}r die zivile Sicherheit{\textquoteright} and the project SOLIDE (no. 13N14456). Publisher Copyright: {\textcopyright} 2017 Copyright held by the owner/author(s).; The 9th International Conference on Knowledge Capture - K-CAP 2017 : 9th International Conference on Knowledge Capture, K-CAP 2017 ; Conference date: 04-12-2017 Through 06-12-2017",

year = "2017",

month = dec,

day = "4",

doi = "10.1145/3148011.3148024",

language = "English",

isbn = "978-1-4503-5553-7",

series = "Proceedings of the Knowledge Capture Conference, K-CAP 2017",

publisher = "Association for Computing Machinery, Inc",

pages = "1--8",

booktitle = "Proceedings of the Knowledge Capture Conference, K-CAP 2017",

address = "United States",

url = "http://www.k-cap2017.org, https://k-cap2017.org",

}

Moussallem, D, Usbeck, R, Röder, M & Ngomo, ACN 2017, MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach. in Proceedings of the Knowledge Capture Conference, K-CAP 2017., 9, Proceedings of the Knowledge Capture Conference, K-CAP 2017, Association for Computing Machinery, Inc, S. 1-8, The 9th International Conference on Knowledge Capture - K-CAP 2017, Austin, Texas, USA / Vereinigte Staaten, 04.12.17. https://doi.org/10.1145/3148011.3148024, https://doi.org/10.48550/arXiv.1707.05288

MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach. / Moussallem, Diego; Usbeck, Ricardo; Röder, Michael et al.
Proceedings of the Knowledge Capture Conference, K-CAP 2017. Association for Computing Machinery, Inc, 2017. S. 1-8 9 (Proceedings of the Knowledge Capture Conference, K-CAP 2017).

Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › Begutachtung

TY - CHAP

T1 - MAG

T2 - The 9th International Conference on Knowledge Capture - K-CAP 2017

AU - Moussallem, Diego

AU - Usbeck, Ricardo

AU - Röder, Michael

AU - Ngomo, Axel Cyrille Ngonga

N1 - Conference code: 9

PY - 2017/12/4

Y1 - 2017/12/4

N2 - Entity linking has recently been the subject of a significant body of research. Currently, the best performing approaches rely on trained mono-lingual models. Porting these approaches to other languages is consequently a difficult endeavor as it requires corresponding training data and retraining of the models. We address this drawback by presenting a novel multilingual, knowledge-base agnostic and deterministic approach to entity linking, dubbed MAG. MAG is based on a combination of context-based retrieval on structured knowledge bases and graph algorithms. We evaluate MAG on 23 data sets and in 7 languages. Our results showthat the best approach trained on English datasets (PBOH) achieves a micro F-measure that is up to 4 times worse on datasets in other languages. MAG on the other hand achieves state-of-the-art performance on English datasets and reaches a micro F-measure that is up to 0.6 higher than that of PBOH on non-English languages.

AB - Entity linking has recently been the subject of a significant body of research. Currently, the best performing approaches rely on trained mono-lingual models. Porting these approaches to other languages is consequently a difficult endeavor as it requires corresponding training data and retraining of the models. We address this drawback by presenting a novel multilingual, knowledge-base agnostic and deterministic approach to entity linking, dubbed MAG. MAG is based on a combination of context-based retrieval on structured knowledge bases and graph algorithms. We evaluate MAG on 23 data sets and in 7 languages. Our results showthat the best approach trained on English datasets (PBOH) achieves a micro F-measure that is up to 4 times worse on datasets in other languages. MAG on the other hand achieves state-of-the-art performance on English datasets and reaches a micro F-measure that is up to 0.6 higher than that of PBOH on non-English languages.

KW - Entity Linking

KW - Multilingual

KW - Named Entity Disambiguation

KW - Informatics

KW - Business informatics

UR - https://www.scopus.com/pages/publications/85040571618

UR - https://www.mendeley.com/catalogue/e415dc14-00af-324d-b2cb-3106fc863441/

U2 - 10.1145/3148011.3148024

DO - 10.1145/3148011.3148024

M3 - Article in conference proceedings

AN - SCOPUS:85040571618

SN - 978-1-4503-5553-7

T3 - Proceedings of the Knowledge Capture Conference, K-CAP 2017

SP - 1

EP - 8

BT - Proceedings of the Knowledge Capture Conference, K-CAP 2017

PB - Association for Computing Machinery, Inc

Y2 - 4 December 2017 through 6 December 2017

ER -

Moussallem D, Usbeck R, Röder M, Ngomo ACN. MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach. in Proceedings of the Knowledge Capture Conference, K-CAP 2017. Association for Computing Machinery, Inc. 2017. S. 1-8. 9. (Proceedings of the Knowledge Capture Conference, K-CAP 2017). doi: 10.1145/3148011.3148024, 10.48550/arXiv.1707.05288

MAG: A multilingual, knowledge-base agnostic and deterministic entity linking approach

Abstract

Fachgebiete und Schlagwörter

Zugriff auf Dokument

Fingerprint

Dieses zitieren