Relation classification using coarse and fine-grained networks with SDP supervised key words selection

Yiping Sun, Yu Cui, Takayuki Furuzuki, Weijia Jia

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In relation classification, previous work focused on either whole sentence or key words, meeting problems when sentence contains noise or key words are extracted falsely. In this paper, we propose coarse and fine-grained networks for relation classification, which combine sentence and key words together to be more robust. Then, we propose a word selection network under shortest dependency path (SDP) supervision to select key words automatically instead of pre-processed key words and attention, which guides word selection network to a better feature space. A novel opposite loss is also proposed by pushing useful information in unselected words back to selected ones. In SemEval-2010 Task 8, results show that under the same features, proposed method outperforms state-of-the-art methods for relation classification.

Original languageEnglish
Title of host publicationKnowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings
EditorsWeiru Liu, Bo Yang, Fausto Giunchiglia
PublisherSpringer-Verlag
Pages514-522
Number of pages9
ISBN (Print)9783319993645
DOIs
Publication statusPublished - 2018 Jan 1
Event11th International Conference on Knowledge Science, Engineering and Management, KSEM 2018 - Changchun, China
Duration: 2018 Aug 172018 Aug 19

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11061 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other11th International Conference on Knowledge Science, Engineering and Management, KSEM 2018
CountryChina
CityChangchun
Period18/8/1718/8/19

Fingerprint

Path
Word problem
Feature Space

Keywords

  • Coarse and fine-grained networks
  • Opposite loss
  • Relation classification
  • selection
  • Shortest dependency path

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Sun, Y., Cui, Y., Furuzuki, T., & Jia, W. (2018). Relation classification using coarse and fine-grained networks with SDP supervised key words selection. In W. Liu, B. Yang, & F. Giunchiglia (Eds.), Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings (pp. 514-522). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11061 LNAI). Springer-Verlag. https://doi.org/10.1007/978-3-319-99365-2_46

Relation classification using coarse and fine-grained networks with SDP supervised key words selection. / Sun, Yiping; Cui, Yu; Furuzuki, Takayuki; Jia, Weijia.

Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings. ed. / Weiru Liu; Bo Yang; Fausto Giunchiglia. Springer-Verlag, 2018. p. 514-522 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11061 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sun, Y, Cui, Y, Furuzuki, T & Jia, W 2018, Relation classification using coarse and fine-grained networks with SDP supervised key words selection. in W Liu, B Yang & F Giunchiglia (eds), Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11061 LNAI, Springer-Verlag, pp. 514-522, 11th International Conference on Knowledge Science, Engineering and Management, KSEM 2018, Changchun, China, 18/8/17. https://doi.org/10.1007/978-3-319-99365-2_46
Sun Y, Cui Y, Furuzuki T, Jia W. Relation classification using coarse and fine-grained networks with SDP supervised key words selection. In Liu W, Yang B, Giunchiglia F, editors, Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings. Springer-Verlag. 2018. p. 514-522. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-99365-2_46
Sun, Yiping ; Cui, Yu ; Furuzuki, Takayuki ; Jia, Weijia. / Relation classification using coarse and fine-grained networks with SDP supervised key words selection. Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings. editor / Weiru Liu ; Bo Yang ; Fausto Giunchiglia. Springer-Verlag, 2018. pp. 514-522 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d4171777579b46ed8dad9a4cbcf009de,
title = "Relation classification using coarse and fine-grained networks with SDP supervised key words selection",
abstract = "In relation classification, previous work focused on either whole sentence or key words, meeting problems when sentence contains noise or key words are extracted falsely. In this paper, we propose coarse and fine-grained networks for relation classification, which combine sentence and key words together to be more robust. Then, we propose a word selection network under shortest dependency path (SDP) supervision to select key words automatically instead of pre-processed key words and attention, which guides word selection network to a better feature space. A novel opposite loss is also proposed by pushing useful information in unselected words back to selected ones. In SemEval-2010 Task 8, results show that under the same features, proposed method outperforms state-of-the-art methods for relation classification.",
keywords = "Coarse and fine-grained networks, Opposite loss, Relation classification, selection, Shortest dependency path",
author = "Yiping Sun and Yu Cui and Takayuki Furuzuki and Weijia Jia",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-99365-2_46",
language = "English",
isbn = "9783319993645",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer-Verlag",
pages = "514--522",
editor = "Weiru Liu and Bo Yang and Fausto Giunchiglia",
booktitle = "Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings",

}

TY - GEN

T1 - Relation classification using coarse and fine-grained networks with SDP supervised key words selection

AU - Sun, Yiping

AU - Cui, Yu

AU - Furuzuki, Takayuki

AU - Jia, Weijia

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In relation classification, previous work focused on either whole sentence or key words, meeting problems when sentence contains noise or key words are extracted falsely. In this paper, we propose coarse and fine-grained networks for relation classification, which combine sentence and key words together to be more robust. Then, we propose a word selection network under shortest dependency path (SDP) supervision to select key words automatically instead of pre-processed key words and attention, which guides word selection network to a better feature space. A novel opposite loss is also proposed by pushing useful information in unselected words back to selected ones. In SemEval-2010 Task 8, results show that under the same features, proposed method outperforms state-of-the-art methods for relation classification.

AB - In relation classification, previous work focused on either whole sentence or key words, meeting problems when sentence contains noise or key words are extracted falsely. In this paper, we propose coarse and fine-grained networks for relation classification, which combine sentence and key words together to be more robust. Then, we propose a word selection network under shortest dependency path (SDP) supervision to select key words automatically instead of pre-processed key words and attention, which guides word selection network to a better feature space. A novel opposite loss is also proposed by pushing useful information in unselected words back to selected ones. In SemEval-2010 Task 8, results show that under the same features, proposed method outperforms state-of-the-art methods for relation classification.

KW - Coarse and fine-grained networks

KW - Opposite loss

KW - Relation classification

KW - selection

KW - Shortest dependency path

UR - http://www.scopus.com/inward/record.url?scp=85052229682&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052229682&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-99365-2_46

DO - 10.1007/978-3-319-99365-2_46

M3 - Conference contribution

SN - 9783319993645

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

SP - 514

EP - 522

BT - Knowledge Science, Engineering and Management - 11th International Conference, KSEM 2018, Proceedings

A2 - Liu, Weiru

A2 - Yang, Bo

A2 - Giunchiglia, Fausto

PB - Springer-Verlag

ER -