Title:
NETWORK TRAINING METHOD AND APPARATUS, ROBOT CONTROL METHOD AND APPARATUS, DEVICE, STORAGE MEDIUM, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2023/123838
Kind Code:
A1
Abstract:
A network training method and apparatus, a robot control method and apparatus, an electronic device, a computer storage medium, and a computer program. The training method comprises: acquiring environmental state information in a target application scenario (S101); according to the environmental state information and a pre-trained reinforcement learning network, obtaining action sequence information, and determining a total reward return value corresponding to the action sequence information, the action sequence information being used for indicating at least two continuous execution actions within a future preset duration (S102); and on the basis of the total reward return value, adjusting a network parameter value of the reinforcement learning network so as to obtain a trained reinforcement learning network (S103).
Inventors:
LI CHUMING (CN)
LIU YU (CN)
WANG XIAOGANG (CN)
LIU YU (CN)
WANG XIAOGANG (CN)
Application Number:
PCT/CN2022/094863
Publication Date:
July 06, 2023
Filing Date:
May 25, 2022
Export Citation:
Assignee:
SHANGHAI SENSETIME INTELLIGENT TECH CO LTD (CN)
International Classes:
G05B13/04; G06N3/04; G06N3/08
Foreign References:
CN114397817A | 2022-04-26 | |||
CN111612126A | 2020-09-01 | |||
CN113156892A | 2021-07-23 | |||
CN113326872A | 2021-08-31 | |||
US20210397961A1 | 2021-12-23 | |||
CN112882469A | 2021-06-01 | |||
US20210247744A1 | 2021-08-12 | |||
CN111602144A | 2020-08-28 | |||
CN113077052A | 2021-07-06 |
Attorney, Agent or Firm:
CHINA PAT INTELLECTUAL PROPERTY OFFICE (CN)
Download PDF: