Skip to main navigation Skip to search Skip to main content

Enriching BERT With Knowledge Graph Embedding For Industry Classification

  • Shiyue Wang
  • , Youcheng Pan
  • , Zhenran Xu
  • , Baotian Hu*
  • , Xiaolong Wang
  • *Corresponding author for this work
  • Harbin Institute of Technology Shenzhen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Industry classification for startup companies is meaningful not only to navigate investment strategies but also to find potential competitors. It is essentially a challenging domain-specific text classification task. Due to the lack of such dataset, in this paper, we first construct a dataset for industry classification based on the companies listed on the Chinese National Equities Exchange and Quotations (NEEQ), which consists of 17, 604 annual business reports and their corresponding industry labels. Second, we introduce a novel Knowledge Graph Enriched BERT model (KGEB), which can understand a domain-specific text by enhancing the word representation with external knowledge and can take full use of the local knowledge graph without pre-training. Experimental results show the promising performance of the proposed model and demonstrate its effectiveness for tackling the domain-specific classification task.

Original languageEnglish
Title of host publicationNeural Information Processing - 28th International Conference, ICONIP 2021, Proceedings
EditorsTeddy Mantoro, Minho Lee, Media Anugerah Ayu, Kok Wai Wong, Achmad Nizar Hidayanto
PublisherSpringer Science and Business Media Deutschland GmbH
Pages709-717
Number of pages9
ISBN (Print)9783030923099
DOIs
StatePublished - 2021
Externally publishedYes
Event28th International Conference on Neural Information Processing, ICONIP 2021 - Virtual, Online, Indonesia
Duration: 8 Dec 202112 Dec 2021

Publication series

NameCommunications in Computer and Information Science
Volume1517 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference28th International Conference on Neural Information Processing, ICONIP 2021
Country/TerritoryIndonesia
CityVirtual, Online
Period8/12/2112/12/21

Keywords

  • Graph convolutional network
  • Industry classification
  • Knowledge graph

Fingerprint

Dive into the research topics of 'Enriching BERT With Knowledge Graph Embedding For Industry Classification'. Together they form a unique fingerprint.

Cite this