PNEL: Pointer Network Based End-To-End Entity Linking over Knowledge Graphs
Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Standard
PNEL : Pointer Network Based End-To-End Entity Linking over Knowledge Graphs. / Banerjee, Debayan; Chaudhuri, Debanjan; Dubey, Mohnish et al.
The Semantic Web – ISWC 2020 - 19th International Semantic Web Conference, 2020, Proceedings. ed. / Jeff Z. Pan; Valentina Tamma; Claudia d’Amato; Krzysztof Janowicz; Bo Fu; Axel Polleres; Oshani Seneviratne; Lalana Kagal. Vol. 1 Cham : Springer Science and Business Media Deutschland GmbH, 2020. p. 21-38 (Lecture Notes in Computer Science ; Vol. 12506 LNCS).Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - CHAP
T1 - PNEL
T2 - 19th International Semantic Web Conference - ISWC 2020
AU - Banerjee, Debayan
AU - Chaudhuri, Debanjan
AU - Dubey, Mohnish
AU - Lehmann, Jens
N1 - Conference code: 19
PY - 2020
Y1 - 2020
N2 - Question Answering systems are generally modelled as a pipeline consisting of a sequence of steps. In such a pipeline, Entity Linking (EL) is often the first step. Several EL models first perform span detection and then entity disambiguation. In such models errors from the span detection phase cascade to later steps and result in a drop of overall accuracy. Moreover, lack of gold entity spans in training data is a limiting factor for span detector training. Hence the movement towards end-to-end EL models began where no separate span detection step is involved. In this work we present a novel approach to end-to-end EL by applying the popular Pointer Network model, which achieves competitive performance. We demonstrate this in our evaluation over three datasets on the Wikidata Knowledge Graph.
AB - Question Answering systems are generally modelled as a pipeline consisting of a sequence of steps. In such a pipeline, Entity Linking (EL) is often the first step. Several EL models first perform span detection and then entity disambiguation. In such models errors from the span detection phase cascade to later steps and result in a drop of overall accuracy. Moreover, lack of gold entity spans in training data is a limiting factor for span detector training. Hence the movement towards end-to-end EL models began where no separate span detection step is involved. In this work we present a novel approach to end-to-end EL by applying the popular Pointer Network model, which achieves competitive performance. We demonstrate this in our evaluation over three datasets on the Wikidata Knowledge Graph.
KW - Entity Linking
KW - Knowledge Graphs
KW - Question Answering
KW - Wikidata
KW - Informatics
UR - http://www.scopus.com/inward/record.url?scp=85096601166&partnerID=8YFLogxK
UR - https://d-nb.info/1220608777
UR - https://www.mendeley.com/catalogue/650bc4a3-bf2f-36ef-8584-1e4c208459ad/
U2 - 10.1007/978-3-030-62419-4_2
DO - 10.1007/978-3-030-62419-4_2
M3 - Article in conference proceedings
AN - SCOPUS:85096601166
SN - 978-3-030-62418-7
VL - 1
T3 - Lecture Notes in Computer Science
SP - 21
EP - 38
BT - The Semantic Web – ISWC 2020 - 19th International Semantic Web Conference, 2020, Proceedings
A2 - Pan, Jeff Z.
A2 - Tamma, Valentina
A2 - d’Amato, Claudia
A2 - Janowicz, Krzysztof
A2 - Fu, Bo
A2 - Polleres, Axel
A2 - Seneviratne, Oshani
A2 - Kagal, Lalana
PB - Springer Science and Business Media Deutschland GmbH
CY - Cham
Y2 - 1 November 2020 through 6 November 2020
ER -