Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
FEDERATED LEARNING METHOD AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2023/029944
Kind Code:
A1
Abstract:
A federated learning method and device. The method comprises: obtaining N operators from an operator library according to a target federated learning task (S110), the operator library comprising K operators associated with federated learning, N and K being an integer greater than 1, and N being less than or equal to K; compounding multiple operators in the N operators to obtain a first composite operator (S120); processing a first dataset by means of the first composite operator to obtain a second dataset (S130), the first dataset being a training set for a target federated learning; and using the second dataset to perform federated learning to obtain a first federated learning model (S140). In the method, by means of a cascade meta-learning structure, an output method combination structure can be adapted to the problem of optimizing an actual scene of federated learning, so as to improve the efficiency and quality of federated learning.

Inventors:
ZHENG ZIMU (CN)
LI RAN (CN)
SUN MING (CN)
SHAN YUXIN (CN)
PU JIE (CN)
MA JIANWEI (CN)
Application Number:
PCT/CN2022/112055
Publication Date:
March 09, 2023
Filing Date:
August 12, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
HUAWEI CLOUD COMPUTING TECH CO LTD (CN)
International Classes:
G06N20/00
Domestic Patent References:
WO2021158313A12021-08-12
Foreign References:
CN112733967A2021-04-30
CN113128419A2021-07-16
CN111310934A2020-06-19
CN113052334A2021-06-29
CN112464287A2021-03-09
CN112598121A2021-04-02
CN112257876A2021-01-22
Attorney, Agent or Firm:
LONGSUN LEAD IP LTD. (CN)
Download PDF: