ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training

Hongyu Zhou; Chengjie Sun; Lei Lin; Lili Shan

doi:10.1007/978-3-031-44699-3_5

ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training

Hongyu Zhou
, Chengjie Sun^*
, Lei Lin
, Lili Shan

^*Corresponding author for this work

School of Computer Science and Technology, Harbin Institute of Technology

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

4 Scopus citations

Abstract

This article proposes a Chinese few-shot emerging entity linking model based on ERNIE and adversarial training. The model utilizes ERNIE as the base model and achieves accurate linking of Chinese few-shot emerging entities by adding adversarial perturbations during the training process. Experimental results on standard entity linking evaluation datasets demonstrate significant performance improvements of our proposed model compared to baseline models. Moreover, we compare multiple candidate entity retrieval methods through comparative experiments to evaluate and compare their effectiveness in the entity linking task. The experimental results show that our appoarch achieved an F1 score of 0.60 and ranked second in the NLPCC 2023 Shared Task 6 (Chinese Few-shot and Zero-shot Entity Linking). Experimental results demonstrate that the model has high predictive performance and robustness in the Chinese few-shot emerging entity linking task, providing reference and inspiration for research and practice in related fields.

Original language	English
Title of host publication	Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings
Editors	Fei Liu, Nan Duan, Qingting Xu, Yu Hong
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	48-56
Number of pages	9
ISBN (Print)	9783031446986
DOIs	https://doi.org/10.1007/978-3-031-44699-3_5
State	Published - 2023
Externally published	Yes
Event	12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023 - Foshan, China Duration: 12 Oct 2023 → 15 Oct 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14304 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023
Country/Territory	China
City	Foshan
Period	12/10/23 → 15/10/23

Keywords

Adversarial Training
Candidate Entity Retrieval
Entity Linking

Access to Document

10.1007/978-3-031-44699-3_5

Cite this

Zhou, H., Sun, C., Lin, L., & Shan, L. (2023). ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training. In F. Liu, N. Duan, Q. Xu, & Y. Hong (Eds.), Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings (pp. 48-56). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14304 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-44699-3_5

Zhou, Hongyu ; Sun, Chengjie ; Lin, Lei et al. / ERNIE-AT-CEL : A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training. Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings. editor / Fei Liu ; Nan Duan ; Qingting Xu ; Yu Hong. Springer Science and Business Media Deutschland GmbH, 2023. pp. 48-56 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{75e3c49a726446ef90f4272ed55a6207,

title = "ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training",

abstract = "This article proposes a Chinese few-shot emerging entity linking model based on ERNIE and adversarial training. The model utilizes ERNIE as the base model and achieves accurate linking of Chinese few-shot emerging entities by adding adversarial perturbations during the training process. Experimental results on standard entity linking evaluation datasets demonstrate significant performance improvements of our proposed model compared to baseline models. Moreover, we compare multiple candidate entity retrieval methods through comparative experiments to evaluate and compare their effectiveness in the entity linking task. The experimental results show that our appoarch achieved an F1 score of 0.60 and ranked second in the NLPCC 2023 Shared Task 6 (Chinese Few-shot and Zero-shot Entity Linking). Experimental results demonstrate that the model has high predictive performance and robustness in the Chinese few-shot emerging entity linking task, providing reference and inspiration for research and practice in related fields.",

keywords = "Adversarial Training, Candidate Entity Retrieval, Entity Linking",

author = "Hongyu Zhou and Chengjie Sun and Lei Lin and Lili Shan",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023 ; Conference date: 12-10-2023 Through 15-10-2023",

year = "2023",

doi = "10.1007/978-3-031-44699-3\_5",

language = "英语",

isbn = "9783031446986",

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 = "48--56",

editor = "Fei Liu and Nan Duan and Qingting Xu and Yu Hong",

booktitle = "Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings",

address = "德国",

}

Zhou, H, Sun, C , Lin, L & Shan, L 2023, ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training. in F Liu, N Duan, Q Xu & Y Hong (eds), Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14304 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 48-56, 12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023, Foshan, China, 12/10/23. https://doi.org/10.1007/978-3-031-44699-3_5

ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training. / Zhou, Hongyu; Sun, Chengjie ; Lin, Lei et al.
Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings. ed. / Fei Liu; Nan Duan; Qingting Xu; Yu Hong. Springer Science and Business Media Deutschland GmbH, 2023. p. 48-56 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14304 LNAI).

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

TY - GEN

T1 - ERNIE-AT-CEL

T2 - 12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023

AU - Zhou, Hongyu

AU - Sun, Chengjie

AU - Lin, Lei

AU - Shan, Lili

PY - 2023

Y1 - 2023

N2 - This article proposes a Chinese few-shot emerging entity linking model based on ERNIE and adversarial training. The model utilizes ERNIE as the base model and achieves accurate linking of Chinese few-shot emerging entities by adding adversarial perturbations during the training process. Experimental results on standard entity linking evaluation datasets demonstrate significant performance improvements of our proposed model compared to baseline models. Moreover, we compare multiple candidate entity retrieval methods through comparative experiments to evaluate and compare their effectiveness in the entity linking task. The experimental results show that our appoarch achieved an F1 score of 0.60 and ranked second in the NLPCC 2023 Shared Task 6 (Chinese Few-shot and Zero-shot Entity Linking). Experimental results demonstrate that the model has high predictive performance and robustness in the Chinese few-shot emerging entity linking task, providing reference and inspiration for research and practice in related fields.

AB - This article proposes a Chinese few-shot emerging entity linking model based on ERNIE and adversarial training. The model utilizes ERNIE as the base model and achieves accurate linking of Chinese few-shot emerging entities by adding adversarial perturbations during the training process. Experimental results on standard entity linking evaluation datasets demonstrate significant performance improvements of our proposed model compared to baseline models. Moreover, we compare multiple candidate entity retrieval methods through comparative experiments to evaluate and compare their effectiveness in the entity linking task. The experimental results show that our appoarch achieved an F1 score of 0.60 and ranked second in the NLPCC 2023 Shared Task 6 (Chinese Few-shot and Zero-shot Entity Linking). Experimental results demonstrate that the model has high predictive performance and robustness in the Chinese few-shot emerging entity linking task, providing reference and inspiration for research and practice in related fields.

KW - Adversarial Training

KW - Candidate Entity Retrieval

KW - Entity Linking

UR - https://www.scopus.com/pages/publications/85174504493

U2 - 10.1007/978-3-031-44699-3_5

DO - 10.1007/978-3-031-44699-3_5

M3 - 会议稿件

AN - SCOPUS:85174504493

SN - 9783031446986

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

SP - 48

EP - 56

BT - Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings

A2 - Liu, Fei

A2 - Duan, Nan

A2 - Xu, Qingting

A2 - Hong, Yu

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 12 October 2023 through 15 October 2023

ER -

Zhou H, Sun C , Lin L , Shan L. ERNIE-AT-CEL: A Chinese Few-Shot Emerging Entity Linking Model Based on ERNIE and Adversarial Training. In Liu F, Duan N, Xu Q, Hong Y, editors, Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 48-56. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-44699-3_5