EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Identify novel elements of knowledge with word embedding

Deyun Yin iD, Zhao Wu, Kazuki Yokota, Kuniko Matsumoto, Sotaro Shibayama iD

PLOS ONE · 20.06.2023 · 18 (6) · e0284567

DOI: 10.1371/journal.pone.0284567 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 20.06.2023. Verfasst von Deyun Yin und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Economics and Management, Harbin Institute of Technology (Shenzhen), Shenzhen, China. Hauptthemen laut MEDLINE: Semantics, Machine Learning. Die Arbeit wurde bisher 29-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Lars Erik Lundbergs Stiftelse för Forskning och Utbildning, Swedish Foundation for International Cooperation in Research and Higher Education, Japan Society for Aeronautical and Space Sciences.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As novelty is a core value in science, a reliable approach to measuring the novelty of scientific documents is critical. Previous novelty measures however had a few limitations. First, the majority of previous measures are based on recombinant novelty concept, attempting to identify a novel combination of knowledge elements, but insufficient effort has been made to identify a novel element itself ( element novelty ). Second, most previous measures are not validated, and it is unclear what aspect of newness is measured. Third, some of the previous measures can be computed only in certain scientific fields for technical constraints. This study thus aims to provide a validated and field-universal approach to computing element novelty. We drew on machine learning to develop a word embedding model, which allows us to extract semantic information from text data. Our validation analyses suggest that our word embedding model does convey semantic information. Based on the trained word embedding, we quantified the element novelty of a document by measuring its distance from the rest of the document universe. We then carried out a questionnaire survey to obtain self-reported novelty scores from 800 scientists. We found that our element novelty measure is significantly correlated with self-reported novelty in terms of discovering and identifying new phenomena, substances, molecules, etc. and that this correlation is observed across different scientific fields.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

SemanticsMachine LearningHumansSurveys and QuestionnairesSelf Report

Autor:innen und Institutionen

  • School of Economics and Management, Harbin Institute of Technology (Shenzhen), Shenzhen, China.
  • World Intellectual Property Organization, Geneva, Switzerland.
  • School of Business Administration, Hitotsubashi University, Tokyo, Japan.
  • National Institute of Science and Technology Policy, Tokyo, Japan.
  • CIRCLE, Lund University, Lund, Sweden.
  • Institute for Future Initiative, The University of Tokyo, Tokyo, Japan.

ORCID-Profile: Deyun Yin, Sotaro Shibayama

Förderung

  • Lars Erik Lundbergs Stiftelse för Forskning och Utbildning
  • Swedish Foundation for International Cooperation in Research and Higher Education
  • Japan Society for Aeronautical and Space Sciences
  • Scientific Research Funding for Overseas High-Caliber Personnel of Shenzhen

Zitieren

APA

Yin, D., Wu, Z., Yokota, K., Matsumoto, K., & Shibayama, S. (2023). Identify novel elements of knowledge with word embedding. PLOS ONE, 18(6), e0284567. https://doi.org/10.1371/journal.pone.0284567

Vancouver

Yin D, Wu Z, Yokota K, Matsumoto K, Shibayama S. Identify novel elements of knowledge with word embedding. PLOS ONE. 2023;18(6):e0284567. doi:10.1371/journal.pone.0284567

BibTeX
@article{Yin2023,
  author = {Yin, Deyun and Wu, Zhao and Yokota, Kazuki and Matsumoto, Kuniko and Shibayama, Sotaro},
  title = {Identify novel elements of knowledge with word embedding},
  journal = {PLOS ONE},
  year = {2023},
  volume = {18},
  number = {6},
  pages = {e0284567},
  doi = {10.1371/journal.pone.0284567}
}
Kostenlos starten →