Title:
PREDICTION METHOD AND APPARATUS IN JOINT LEARNING
Document Type and Number:
WIPO Patent Application WO/2023/124312
Kind Code:
A1
Abstract:
A prediction method and apparatus in joint learning. The method comprises: using a participant model of a plurality of participants to predict data to be predicted to obtain a first prediction result (S201); updating model parameters of the participant model of the plurality of participants on the basis of the first prediction result (S202); re-predicting the data to be predicted by using the participant model of the plurality of participants after updating the model parameters to obtain a second prediction result (S203); and calculating a third prediction result according to the second prediction result corresponding to the plurality of participants (S204). The method solves the problems that a prediction result and model training are not associated in joint learning and that only a joint learning model can be used for prediction.
Inventors:
ZHANG MIN (CN)
WANG RUIYANG (CN)
WANG RUIYANG (CN)
Application Number:
PCT/CN2022/121483
Publication Date:
July 06, 2023
Filing Date:
September 26, 2022
Export Citation:
Assignee:
ENNEW ICOME INTERNET TECH CO LTD (CN)
International Classes:
G06N20/00
Foreign References:
CN110929887A | 2020-03-27 | |||
CN112381236A | 2021-02-19 | |||
CN113221183A | 2021-08-06 | |||
CN112948885A | 2021-06-11 | |||
CN113642809A | 2021-11-12 |
Attorney, Agent or Firm:
BEIJING JIAKE INTELLECTUAL PROPERTY LAW FIRM (CN)
Download PDF: