Zero-Shot Embedding for Unseen Entities in Knowledge Graph

Yu ZHAO; Sheng GAO; Patrick GALLINARI; Jun GUO

doi:10.1587/transinf.2016EDP7446

Zero-Shot Embedding for Unseen Entities in Knowledge Graph

Yu ZHAO, Sheng GAO, Patrick GALLINARI, Jun GUO

Full Text Views

0

Cite this

Summary :

Knowledge graph (KG) embedding aims at learning the latent semantic representations for entities and relations. However, most existing approaches can only be applied to KG completion, so cannot identify relations including unseen entities (or Out-of-KG entities). In this paper, motivated by the zero-shot learning, we propose a novel model, namely JointE, jointly learning KG and entity descriptions embedding, to extend KG by adding new relations with Out-of-KG entities. The JointE model is evaluated on entity prediction for zero-shot embedding. Empirical comparisons on benchmark datasets show that the proposed JointE model outperforms state-of-the-art approaches. The source code of JointE is available at https://github.com/yzur/JointE.

Publication: IEICE TRANSACTIONS on Information Vol.E100-D No.7 pp.1440-1447

Publication Date: 2017/07/01

Publicized: 2017/04/10

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2016EDP7446

Type of Manuscript: PAPER

Category: Artificial Intelligence, Data Mining

Authors

Yu ZHAO
  Beijing University of Posts and Telecommunications
Sheng GAO
  Beijing University of Posts and Telecommunications
Patrick GALLINARI
  LIP6, Universit Pierre et Marie Curie
Jun GUO
  Beijing University of Posts and Telecommunications

Keyword

zero-shot learning, knowledge graph, embedding learning, relation prediction

Cite this

Copy

Yu ZHAO, Sheng GAO, Patrick GALLINARI, Jun GUO, "Zero-Shot Embedding for Unseen Entities in Knowledge Graph" in IEICE TRANSACTIONS on Information, vol. E100-D, no. 7, pp. 1440-1447, July 2017, doi: 10.1587/transinf.2016EDP7446.
Abstract: Knowledge graph (KG) embedding aims at learning the latent semantic representations for entities and relations. However, most existing approaches can only be applied to KG completion, so cannot identify relations including unseen entities (or Out-of-KG entities). In this paper, motivated by the zero-shot learning, we propose a novel model, namely JointE, jointly learning KG and entity descriptions embedding, to extend KG by adding new relations with Out-of-KG entities. The JointE model is evaluated on entity prediction for zero-shot embedding. Empirical comparisons on benchmark datasets show that the proposed JointE model outperforms state-of-the-art approaches. The source code of JointE is available at https://github.com/yzur/JointE.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2016EDP7446/_p

Copy

@ARTICLE{e100-d_7_1440,
author={Yu ZHAO, Sheng GAO, Patrick GALLINARI, Jun GUO, },
journal={IEICE TRANSACTIONS on Information},
title={Zero-Shot Embedding for Unseen Entities in Knowledge Graph},
year={2017},
volume={E100-D},
number={7},
pages={1440-1447},
abstract={Knowledge graph (KG) embedding aims at learning the latent semantic representations for entities and relations. However, most existing approaches can only be applied to KG completion, so cannot identify relations including unseen entities (or Out-of-KG entities). In this paper, motivated by the zero-shot learning, we propose a novel model, namely JointE, jointly learning KG and entity descriptions embedding, to extend KG by adding new relations with Out-of-KG entities. The JointE model is evaluated on entity prediction for zero-shot embedding. Empirical comparisons on benchmark datasets show that the proposed JointE model outperforms state-of-the-art approaches. The source code of JointE is available at https://github.com/yzur/JointE.},
keywords={},
doi={10.1587/transinf.2016EDP7446},
ISSN={1745-1361},
month={July},}

Copy

TY - JOUR
TI - Zero-Shot Embedding for Unseen Entities in Knowledge Graph
T2 - IEICE TRANSACTIONS on Information
SP - 1440
EP - 1447
AU - Yu ZHAO
AU - Sheng GAO
AU - Patrick GALLINARI
AU - Jun GUO
PY - 2017
DO - 10.1587/transinf.2016EDP7446
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 7
JA - IEICE TRANSACTIONS on Information
Y1 - July 2017
AB - Knowledge graph (KG) embedding aims at learning the latent semantic representations for entities and relations. However, most existing approaches can only be applied to KG completion, so cannot identify relations including unseen entities (or Out-of-KG entities). In this paper, motivated by the zero-shot learning, we propose a novel model, namely JointE, jointly learning KG and entity descriptions embedding, to extend KG by adding new relations with Out-of-KG entities. The JointE model is evaluated on entity prediction for zero-shot embedding. Empirical comparisons on benchmark datasets show that the proposed JointE model outperforms state-of-the-art approaches. The source code of JointE is available at https://github.com/yzur/JointE.
ER -