Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network

Xiao Rui Su; Zhu Hong You; Hai Cheng Yi; Bo Wei Zhao

doi:10.1007/978-3-030-84532-2_38

Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network

Xiao Rui Su, Zhu Hong You, Hai Cheng Yi, Bo Wei Zhao

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

3 Scopus citations

Abstract

Drug-drug interaction (DDI) prediction is a challenging problem in drug development and disease treatment. Current computational studies mainly solve this problem by designing features and extracting features from multi-sources. They have limitations in accuracy, knowledge and universality. In this paper, a universal computational method for DDI prediction is proposed. The model is made up of several identical layers. Each layer is composed by a multi-attention unit and a capsule unit. Multi-attention is responsibility for extracting neighborhood features and capturing representation under various semantic relation. Capsule unit is responsibility for features aggregating. Each layer can be viewed as first-order representation. High-order representation is obtained by increasing layers. The structure of whole method is easily understood and implemented. Experiment results show that proposed method can achieve a promising predicting performance and stands out when compared with several popular network embedding methods. Proposed method is a reliable model for DDI prediction and is suitable for other interaction predictions.

Original language	English
Title of host publication	Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings
Editors	De-Shuang Huang, Kang-Hyun Jo, Jianqiang Li, Valeriya Gribova, Vitoantonio Bevilacqua
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	423-432
Number of pages	10
ISBN (Print)	9783030845315
DOIs	https://doi.org/10.1007/978-3-030-84532-2_38
State	Published - 2021
Externally published	Yes
Event	17th International Conference on Intelligent Computing, ICIC 2021 - Shenzhen, China Duration: 12 Aug 2021 → 15 Aug 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12838 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	17th International Conference on Intelligent Computing, ICIC 2021
Country/Territory	China
City	Shenzhen
Period	12/08/21 → 15/08/21

Keywords

Capsule network
Drug-drug interaction
Feature aggregation
Knowledge graph representation learning
Multi-attention

Access to Document

10.1007/978-3-030-84532-2_38

Cite this

Su, X. R., You, Z. H., Yi, H. C., & Zhao, B. W. (2021). Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network. In D.-S. Huang, K.-H. Jo, J. Li, V. Gribova, & V. Bevilacqua (Eds.), Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings (pp. 423-432). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12838 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-84532-2_38

Su, Xiao Rui ; You, Zhu Hong ; Yi, Hai Cheng et al. / Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network. Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings. editor / De-Shuang Huang ; Kang-Hyun Jo ; Jianqiang Li ; Valeriya Gribova ; Vitoantonio Bevilacqua. Springer Science and Business Media Deutschland GmbH, 2021. pp. 423-432 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{70145aec06304d0a9bb26f9358ec0bc6,

title = "Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network",

abstract = "Drug-drug interaction (DDI) prediction is a challenging problem in drug development and disease treatment. Current computational studies mainly solve this problem by designing features and extracting features from multi-sources. They have limitations in accuracy, knowledge and universality. In this paper, a universal computational method for DDI prediction is proposed. The model is made up of several identical layers. Each layer is composed by a multi-attention unit and a capsule unit. Multi-attention is responsibility for extracting neighborhood features and capturing representation under various semantic relation. Capsule unit is responsibility for features aggregating. Each layer can be viewed as first-order representation. High-order representation is obtained by increasing layers. The structure of whole method is easily understood and implemented. Experiment results show that proposed method can achieve a promising predicting performance and stands out when compared with several popular network embedding methods. Proposed method is a reliable model for DDI prediction and is suitable for other interaction predictions.",

keywords = "Capsule network, Drug-drug interaction, Feature aggregation, Knowledge graph representation learning, Multi-attention",

author = "Su, {Xiao Rui} and You, {Zhu Hong} and Yi, {Hai Cheng} and Zhao, {Bo Wei}",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 17th International Conference on Intelligent Computing, ICIC 2021 ; Conference date: 12-08-2021 Through 15-08-2021",

year = "2021",

doi = "10.1007/978-3-030-84532-2_38",

language = "英语",

isbn = "9783030845315",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "423--432",

editor = "De-Shuang Huang and Kang-Hyun Jo and Jianqiang Li and Valeriya Gribova and Vitoantonio Bevilacqua",

booktitle = "Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings",

}

Su, XR, You, ZH , Yi, HC & Zhao, BW 2021, Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network. in D-S Huang, K-H Jo, J Li, V Gribova & V Bevilacqua (eds), Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12838 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 423-432, 17th International Conference on Intelligent Computing, ICIC 2021, Shenzhen, China, 12/08/21. https://doi.org/10.1007/978-3-030-84532-2_38

Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network. / Su, Xiao Rui; You, Zhu Hong ; Yi, Hai Cheng et al.
Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings. ed. / De-Shuang Huang; Kang-Hyun Jo; Jianqiang Li; Valeriya Gribova; Vitoantonio Bevilacqua. Springer Science and Business Media Deutschland GmbH, 2021. p. 423-432 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12838 LNAI).

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

TY - GEN

T1 - Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network

AU - Su, Xiao Rui

AU - You, Zhu Hong

AU - Yi, Hai Cheng

AU - Zhao, Bo Wei

PY - 2021

Y1 - 2021

N2 - Drug-drug interaction (DDI) prediction is a challenging problem in drug development and disease treatment. Current computational studies mainly solve this problem by designing features and extracting features from multi-sources. They have limitations in accuracy, knowledge and universality. In this paper, a universal computational method for DDI prediction is proposed. The model is made up of several identical layers. Each layer is composed by a multi-attention unit and a capsule unit. Multi-attention is responsibility for extracting neighborhood features and capturing representation under various semantic relation. Capsule unit is responsibility for features aggregating. Each layer can be viewed as first-order representation. High-order representation is obtained by increasing layers. The structure of whole method is easily understood and implemented. Experiment results show that proposed method can achieve a promising predicting performance and stands out when compared with several popular network embedding methods. Proposed method is a reliable model for DDI prediction and is suitable for other interaction predictions.

AB - Drug-drug interaction (DDI) prediction is a challenging problem in drug development and disease treatment. Current computational studies mainly solve this problem by designing features and extracting features from multi-sources. They have limitations in accuracy, knowledge and universality. In this paper, a universal computational method for DDI prediction is proposed. The model is made up of several identical layers. Each layer is composed by a multi-attention unit and a capsule unit. Multi-attention is responsibility for extracting neighborhood features and capturing representation under various semantic relation. Capsule unit is responsibility for features aggregating. Each layer can be viewed as first-order representation. High-order representation is obtained by increasing layers. The structure of whole method is easily understood and implemented. Experiment results show that proposed method can achieve a promising predicting performance and stands out when compared with several popular network embedding methods. Proposed method is a reliable model for DDI prediction and is suitable for other interaction predictions.

KW - Capsule network

KW - Drug-drug interaction

KW - Feature aggregation

KW - Knowledge graph representation learning

KW - Multi-attention

UR - http://www.scopus.com/inward/record.url?scp=85113725791&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-84532-2_38

DO - 10.1007/978-3-030-84532-2_38

M3 - 会议稿件

AN - SCOPUS:85113725791

SN - 9783030845315

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 423

EP - 432

BT - Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings

A2 - Huang, De-Shuang

A2 - Jo, Kang-Hyun

A2 - Li, Jianqiang

A2 - Gribova, Valeriya

A2 - Bevilacqua, Vitoantonio

PB - Springer Science and Business Media Deutschland GmbH

T2 - 17th International Conference on Intelligent Computing, ICIC 2021

Y2 - 12 August 2021 through 15 August 2021

ER -

Su XR, You ZH , Yi HC, Zhao BW. Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network. In Huang DS, Jo KH, Li J, Gribova V, Bevilacqua V, editors, Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 423-432. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-84532-2_38

Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this