Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
MACHINE LEARNING METHOD AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2020/096099
Kind Code:
A1
Abstract:
Provided is a machine learning method that may reduce annotation costs and improve performance of a target model. A machine learning method, according to some embodiments of the present disclosure, that is performed in a computing device may comprise the steps of: obtaining a learning dataset of a first model comprising a plurality of data samples for which no label information is given; calculating a misprediction probability of the first model for the plurality of data samples; constructing a first data sample set by selecting at least one data sample from the plurality of data samples on the basis of the calculated misprediction probability; obtaining first label information for the first data sample set; and performing a first learning on the first model by using the first data sample set and the first label information.

Inventors:
YOO DONG GEUN (KR)
PAENG KYUNG HYUN (KR)
PARK SUNG GYUN (KR)
Application Number:
PCT/KR2018/013667
Publication Date:
May 14, 2020
Filing Date:
November 09, 2018
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
LUNIT INC (KR)
International Classes:
G06N99/00
Foreign References:
JP2018155522A2018-10-04
KR101895121B12018-09-04
JPS6362893B21988-12-05
Other References:
REYES, OSCAR ET AL.: "Effective active learning strategy for multi-label learning", ARTICLE IN NEUROCOMPUTING, August 2017 (2017-08-01), XP085254908, Retrieved from the Internet [retrieved on 20190722]
KHODDELTABANDEH, MEHRAN ET AL.: "Active Learning for Structured Prediction from Partially Labeled Data", ARXIV:1706.02342V2, 9 June 2017 (2017-06-09), XP080768309, Retrieved from the Internet [retrieved on 20190722]
Attorney, Agent or Firm:
YOU ME PATENT AND LAW FIRM (KR)
Download PDF: