Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
HORIZONTAL FEDERATED LEARNING MODELING OPTIMIZATION METHOD, DEVICE, MEDIUM AND PROGRAM PRODUCT
Document Type and Number:
WIPO Patent Application WO/2023/024368
Kind Code:
A1
Abstract:
The present application discloses a horizontal federated learning modeling optimization method, a device, a medium, and a program product, applied to a federation participant. The horizontal federated learning modeling optimization method comprises: according to an acquired training task data set and target white box signature, by means of performing iterative training updating on a local neural network model, embedding the target white box signature into a preset signature position in a network parameter of the local neural network model to obtain a signature-embedded local model; sending the signature-embedded local model to a horizontal federated server, so that the horizontal federated server aggregates signature-embedded local models transmitted by each federation participant to obtain an aggregate signature-embedded local model, different federation participants having different preset signature positions; and optimizing the signature-embedded local model on the basis of the aggregate signature-embedded local model to obtain a target signature-embedded model.

Inventors:
FAN LIXIN (CN)
YANG QIANG (CN)
GU HANLIN (CN)
LI BOWEN (CN)
Application Number:
PCT/CN2021/141275
Publication Date:
March 02, 2023
Filing Date:
December 24, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
WEBANK CO LTD (CN)
International Classes:
G06F30/27; G06N20/00; G06Q50/18
Foreign References:
CN113627086A2021-11-09
CN111639759A2020-09-08
CN108712260A2018-10-26
US20200184036A12020-06-11
CN111666987A2020-09-15
CN111241580A2020-06-05
Other References:
FAN LIXIN; NG KAM WOH; CHAN CHEE SENG; YANG QIANG: "DeepIPR: Deep Neural Network Ownership Verification With Passports", IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, IEEE COMPUTER SOCIETY., USA, vol. 44, no. 10, 11 June 2021 (2021-06-11), USA , pages 6122 - 6139, XP011919877, ISSN: 0162-8828, DOI: 10.1109/TPAMI.2021.3088846
Attorney, Agent or Firm:
CENFO INTELLECTUAL PROPERTY AGENCY (CN)
Download PDF:
View/Download PDF      PDF Help



 
Previous Patent: MIRROR DISCHARGE MACHINING OIL AND PREPARATION METHOD THEREFOR

Next Patent: POWER DISTRIBUTION NETWORK POWER SUPPLY RELIABILITY ASSESSMENT METHOD BASED ON DIGITAL TWIN, AND DEV...