Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
LEARNING DEVICE, LEARNING METHOD, AND LEARNING PROGRAM
Document Type and Number:
WIPO Patent Application WO/2023/214584
Kind Code:
A1
Abstract:
Provided is a learning device comprising a processing unit for performing enhancement learning of a learning model of an agent in a competition environment in which agents compete against each other. The learning model includes a hyper parameter. The processing unit executes: a step for evaluating the strengths of a plurality of the agents serving as opponents of the agent serving as a learning object; a step for setting, for the agent serving as the learning object, a competition probability according to the strengths of the agents serving as opponents; a step for setting, on the basis of the competition probability, the agent who will serve as the opponent; and a step for causing the agent set to serve as the opponent to compete, and executing enhancement learning of the agent serving as the learning object.

Inventors:
KATAOKA YUJIRO (JP)
ITO MASAYUKI (JP)
MATSUNAMI NATSUKI (JP)
Application Number:
PCT/JP2023/017149
Publication Date:
November 09, 2023
Filing Date:
May 02, 2023
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
MITSUBISHI HEAVY IND LTD (JP)
International Classes:
G06N20/00
Domestic Patent References:
WO2021038759A12021-03-04
Foreign References:
CN114330754A2022-04-12
CN113282100A2021-08-20
CN112016704A2020-12-01
Attorney, Agent or Firm:
SAKAI INTERNATIONAL PATENT OFFICE (JP)
Download PDF: