Node co-occurrence based graph neural networks for knowledge graph link prediction

Dai Quoc Nguyen; Vinh Tong; Dinh Phung; Dat Quoc Nguyen

doi:10.1145/3488560.3502183

Node co-occurrence based graph neural networks for knowledge graph link prediction

Dai Quoc Nguyen, Vinh Tong, Dinh Phung, Dat Quoc Nguyen

Department of Data Science & AI

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Other › peer-review

26 Citations (Scopus)

Abstract

We introduce a novel embedding model, named NoGE, which aims to integrate co-occurrence among entities and relations into graph neural networks to improve knowledge graph completion (i.e., link prediction). Given a knowledge graph, NoGE constructs a single graph considering entities and relations as individual nodes. NoGE then computes weights for edges among nodes based on the co-occurrence of entities and relations. Next, NoGE proposes Dual Quaternion Graph Neural Networks (DualQGNN) and utilizes DualQGNN to update vector representations for entity and relation nodes. NoGE then adopts a score function to produce the triple scores. Comprehensive experimental results show that NoGE obtains state-of-the-art results on three new and difficult benchmark datasets CoDEx for knowledge graph completion.

Original language	English
Title of host publication	Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining
Editors	Parth Nagarkar
Place of Publication	New York NY USA
Publisher	Association for Computing Machinery (ACM)
Pages	1589-1592
Number of pages	4
ISBN (Electronic)	9781450391320
DOIs	https://doi.org/10.1145/3488560.3502183
Publication status	Published - 2022
Event	ACM International Conference on Web Search and Data Mining 2022 - Online, United States of America Duration: 21 Feb 2022 → 25 Feb 2022 Conference number: 15th

Conference

Conference	ACM International Conference on Web Search and Data Mining 2022
Abbreviated title	WSDM 2022
Country/Territory	United States of America
City	Online
Period	21/02/22 → 25/02/22

Keywords

Graph neural networks
Knowledge graph completion
Knowledge graph embeddings
Quaternion

Access to Document

10.1145/3488560.3502183

Cite this

@inproceedings{12619447d5a247278ad1520c24ea3eb1,

title = "Node co-occurrence based graph neural networks for knowledge graph link prediction",

abstract = "We introduce a novel embedding model, named NoGE, which aims to integrate co-occurrence among entities and relations into graph neural networks to improve knowledge graph completion (i.e., link prediction). Given a knowledge graph, NoGE constructs a single graph considering entities and relations as individual nodes. NoGE then computes weights for edges among nodes based on the co-occurrence of entities and relations. Next, NoGE proposes Dual Quaternion Graph Neural Networks (DualQGNN) and utilizes DualQGNN to update vector representations for entity and relation nodes. NoGE then adopts a score function to produce the triple scores. Comprehensive experimental results show that NoGE obtains state-of-the-art results on three new and difficult benchmark datasets CoDEx for knowledge graph completion.",

keywords = "Graph neural networks, Knowledge graph completion, Knowledge graph embeddings, Quaternion",

author = "Nguyen, {Dai Quoc} and Vinh Tong and Dinh Phung and Nguyen, {Dat Quoc}",

note = "Publisher Copyright: {\textcopyright} 2022 ACM.; ACM International Conference on Web Search and Data Mining 2022, WSDM 2022 ; Conference date: 21-02-2022 Through 25-02-2022",

year = "2022",

doi = "10.1145/3488560.3502183",

language = "English",

pages = "1589--1592",

editor = "Parth Nagarkar",

booktitle = "Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining",

publisher = "Association for Computing Machinery (ACM)",

address = "United States of America",

}

Nguyen, DQ, Tong, V, Phung, D & Nguyen, DQ 2022, Node co-occurrence based graph neural networks for knowledge graph link prediction. in P Nagarkar (ed.), Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining. Association for Computing Machinery (ACM), New York NY USA, pp. 1589-1592, ACM International Conference on Web Search and Data Mining 2022, Online, United States of America, 21/02/22. https://doi.org/10.1145/3488560.3502183

Node co-occurrence based graph neural networks for knowledge graph link prediction. / Nguyen, Dai Quoc; Tong, Vinh; Phung, Dinh et al.
Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining. ed. / Parth Nagarkar. New York NY USA: Association for Computing Machinery (ACM), 2022. p. 1589-1592.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Other › peer-review

TY - GEN

T1 - Node co-occurrence based graph neural networks for knowledge graph link prediction

AU - Nguyen, Dai Quoc

AU - Tong, Vinh

AU - Phung, Dinh

AU - Nguyen, Dat Quoc

N1 - Conference code: 15th

PY - 2022

Y1 - 2022

N2 - We introduce a novel embedding model, named NoGE, which aims to integrate co-occurrence among entities and relations into graph neural networks to improve knowledge graph completion (i.e., link prediction). Given a knowledge graph, NoGE constructs a single graph considering entities and relations as individual nodes. NoGE then computes weights for edges among nodes based on the co-occurrence of entities and relations. Next, NoGE proposes Dual Quaternion Graph Neural Networks (DualQGNN) and utilizes DualQGNN to update vector representations for entity and relation nodes. NoGE then adopts a score function to produce the triple scores. Comprehensive experimental results show that NoGE obtains state-of-the-art results on three new and difficult benchmark datasets CoDEx for knowledge graph completion.

AB - We introduce a novel embedding model, named NoGE, which aims to integrate co-occurrence among entities and relations into graph neural networks to improve knowledge graph completion (i.e., link prediction). Given a knowledge graph, NoGE constructs a single graph considering entities and relations as individual nodes. NoGE then computes weights for edges among nodes based on the co-occurrence of entities and relations. Next, NoGE proposes Dual Quaternion Graph Neural Networks (DualQGNN) and utilizes DualQGNN to update vector representations for entity and relation nodes. NoGE then adopts a score function to produce the triple scores. Comprehensive experimental results show that NoGE obtains state-of-the-art results on three new and difficult benchmark datasets CoDEx for knowledge graph completion.

KW - Graph neural networks

KW - Knowledge graph completion

KW - Knowledge graph embeddings

KW - Quaternion

UR - http://www.scopus.com/inward/record.url?scp=85125801029&partnerID=8YFLogxK

U2 - 10.1145/3488560.3502183

DO - 10.1145/3488560.3502183

M3 - Conference Paper

AN - SCOPUS:85125801029

SP - 1589

EP - 1592

BT - Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining

A2 - Nagarkar, Parth

PB - Association for Computing Machinery (ACM)

CY - New York NY USA

T2 - ACM International Conference on Web Search and Data Mining 2022

Y2 - 21 February 2022 through 25 February 2022

ER -

Node co-occurrence based graph neural networks for knowledge graph link prediction

Abstract

Conference

Keywords

Access to Document

Other files and links

Cite this