Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
DECENTRALIZED FEDERATED CLUSTERING LEARNING METHOD AND APPARATUS, AND DEVICE AND MEDIUM
Document Type and Number:
WIPO Patent Application WO/2024/082515
Kind Code:
A1
Abstract:
Disclosed in the present application are a decentralized federated clustering learning method and apparatus, and a device and a medium. The method comprises: a task initiator executing, on the basis of a preset clustering algorithm and a preset encryption algorithm, interaction with at least two data parties, so as to determine an optimal initial clustering center and at least two initial clusters; during the process of iterating and updating the optimal initial clustering center and the initial clusters, determining distances from samples in a joint data set to each cluster, acquiring, on the basis of the preset encryption algorithm, the total distance of the samples in the joint data set relative to the current clustering center, and interacting with the at least two data parties according to the total distance, so as to determine the total distance of the three parties; and determining, according to the total distance, whether a preset iteration termination condition is met, and if so, determining that the clustering center after the last iteration and update is a final clustering center. While the privacy and security of a task initiator and data parties are ensured, efficient joint clustering learning can be realized.

Inventors:
SUN YINYIN (CN)
LI ZHONGPING (CN)
Application Number:
PCT/CN2023/079371
Publication Date:
April 25, 2024
Filing Date:
March 02, 2023
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
LINGSHU TECH CO LTD (CN)
International Classes:
G06N20/00
Domestic Patent References:
WO2021249500A12021-12-16
Foreign References:
CN115545215A2022-12-30
CN112231760A2021-01-15
CN114386071A2022-04-22
CN113657525A2021-11-16
Attorney, Agent or Firm:
BEYOND ATTORNEYS AT LAW (CN)
Download PDF: