METHOD FOR OPTIMIZING VERTICAL FEDERATED PREDICTION, DEVICE, MEDIUM, AND COMPUTER PROGRAM PRODUCT

Title:

METHOD FOR OPTIMIZING VERTICAL FEDERATED PREDICTION, DEVICE, MEDIUM, AND COMPUTER PROGRAM PRODUCT

Document Type and Number:

WIPO Patent Application WO/2023/024349

Kind Code:

A1

Abstract:

The present application discloses a method for optimizing vertical federated prediction, a device, a medium, and a computer program product. The method comprises: extracting a sample to be predicted, and, on the basis of a target prediction model from local iterative training, generating a first party model prediction result corresponding to said sample and a corresponding intermediate sample feature; sending the intermediate sample feature to a second device for the second device to, on the basis of a vertical federated residual boosting model, jointly perform model prediction on the intermediate sample feature and an ID matching sample corresponding to the sample to be predicted, obtaining a second party model prediction result; acquiring a first party model weight, and receiving the second party model prediction result and a second party model weight that are sent by the second device; and on the basis of the first party model weight and the second party model weight, carrying out weighted aggregation on the first party model prediction result and the second party model prediction result to obtain a target federated prediction result.

More Like This:

WO/2022/123695	LEARNING DEVICE, SEARCH DEVICE, LEARNING METHOD, SEARCH METHOD, AND PROGRAM
WO/2020/112105	EVENT-BASED PROCESSING USING THE OUTPUT OF A DEEP NEURAL NETWORK
WO/2021/196367	PERFORMANCE-RELATED GROUND MOTION RISK ANALYSIS METHOD EMPLOYING THREE-LAYER DATASET-BASED NEURAL NETWORK

Inventors:

WAN SHENG (CN)
GAO DASHAN (CN)
JU CE (CN)
TAN BEN (CN)
YANG QIANG (CN)

Application Number:

PCT/CN2021/139640

Publication Date:

March 02, 2023

Filing Date:

December 20, 2021

Export Citation:

Click for automatic bibliography generation Help

Assignee:

WEBANK CO LTD (CN)

International Classes:

G06N3/04

Foreign References:

CN113688986A	2021-11-23
CN112785002A	2021-05-11
CN112700010A	2021-04-23
CN109002861A	2018-12-14
CN112001740A	2020-11-27
US20210089964A1	2021-03-25

Attorney, Agent or Firm:

CENFO INTELLECTUAL PROPERTY AGENCY (CN)

Download PDF:

View/Download PDF PDF Help

Previous Patent: RECYCLABLE PRESTRESSED SUPPORT SYSTEM AND CONSTRUCTION METHOD

Next Patent: LONGITUDINAL FEDERAL PREDICTION OPTIMIZATION METHOD AND DEVICE, MEDIUM, AND COMPUTER PROGRAM PRODUCT