Comparative analysis of knowledge bases: ConceptNet vs CYC

Authors

  • Asel Nuraliyeva Kazakh-British Technical University
  • Suleimen Daukishov Kazakh-British Technical University
  • Regina Nassyrova IT Analyst, Philip Morris Kazakhstan

Keywords:

knowledge base, natural language processing, data summarization, question answering, text categorization

Abstract

Data summarization, question answering, text categorization are some of the tasks knowledge bases are used for. A knowledge base (KB) is a computerized compilation of information about the world. They can be useful for complex tasks and problems in NLP and they comprise both entities and relations. We performed an analytical comparison of two knowledge bases which capture a wide variety of common-sense information - ConceptNet and Cyc.

Manually curated knowledge base Cyc has invested more than 1,000 man-years over the last two decades building a knowledge base that is meant to capture a wide range of common-sense skills. On the other hand, ConceptNet is a free multilingual knowledge base and crowdsourced knowledge initiative that uses a large number of links to connect commonplace items. When looking for common sense logic and answers to questions, ConceptNet is a great place to start.

In this research, two well-known knowledge bases were reviewed - ConceptNet and Cyc - their origin, differences, applications, benefits and disadvantages were covered. The authors hope this paper would be useful for researchers looking for more appropriate knowledge bases for word embedding, word sense disambiguation and natural-language communication.

Downloads

Download data is not yet available.

Downloads

Published

2021-05-14

How to Cite

Nuraliyeva, A., Daukishov, S., & Nassyrova, R. (2021). Comparative analysis of knowledge bases: ConceptNet vs CYC . ADVANCED TECHNOLOGIES AND COMPUTER SCIENCE, (1), 4–14. Retrieved from https://atcs.iict.kz/index.php/atcs/article/view/37

Issue

Section

Applied mathematics, computer science and control theory