Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
ELECTRONIC DEVICE FOR DETERMINING SIMILARITY DEGREE AND CONTROL METHOD THEREOF
Document Type and Number:
WIPO Patent Application WO/2021/107360
Kind Code:
A3
Abstract:
An embodiment according to the present invention may provide a control method using machine learning in order to analyze whether graphs are similar, the method comprising the steps of: acquiring node-weight sequences for respective nodes of each of weight graphs; acquiring embedding vectors for the respective nodes by configuring, as an input of a learned auto-encoder model, the node-weight sequences for the respective nodes; acquiring final embedding vectors of each of the weight graphs by using the embedding vectors for the respective nodes; and determining whether the weight graphs are similar, by classifying each of the weight graphs.

Inventors:
LEE KI YONG (KR)
YANG YU JEONG (KR)
SEO MIN JI (KR)
Application Number:
PCT/KR2020/012648
Publication Date:
July 22, 2021
Filing Date:
September 18, 2020
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV SOOKMYUNG WOMENS IND ACAD COOP FOUND (KR)
International Classes:
G06F16/901; G06N20/00
Domestic Patent References:
WO2018058426A12018-04-05
Foreign References:
KR20190109670A2019-09-26
EP3564889A12019-11-06
Other References:
SHIMA KHOSHRAFTAR; SEDIGHEH MAHDAVI; AIJUN AN; YONGGANG HU; JUNFENG LIU: "Dynamic Graph Embedding via LSTM History Tracking", ARXIV.ORG, 5 November 2019 (2019-11-05), pages 1 - 9, XP081525405
YU JIN , JOSEPH F. JAJA: "Learning Graph-Level Representations with Gated Recurrent Neural Networks", ARXIV.ORG, 20 May 2018 (2018-05-20), pages 1 - 10, XP080880199
"Master's thesis", 1 February 2020, SOOKMYUNG WOMEN'S UNIVERSITY GRADUATE SCHOOL : COMPUTER SCIENCE, Korea, article SEO, MIN-JI: "An LSTM Autoencoder-based Embedding Technique for Weighted Graphs", pages: 1 - 80, XP009529262
SEO, MINJI; LEE, KI YONG: "A Weighted Graph Embedding Technique Based on LSTM Autoencoders", PROCEEDINGS OF KOREAN SOFTWARE CONFERENCE 2019 OF THE KOREAN INSTITUTE OF INFORMATION SCIENTISTS AND ENGINEERS, 1 December 2019 (2019-12-01), Korea, pages 464 - 466, XP009529242
Attorney, Agent or Firm:
CHOI, Kwang Seok (KR)
Download PDF: