Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
MODEL TRAINING METHOD AND RELATED SYSTEM, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2022/262557
Kind Code:
A1
Abstract:
Embodiments of the present application provide a model training method and a related system, and a storage medium, which are applied to the technical field of artificial intelligence, for example, federated learning. The system comprises: a client configured to train a first model according to unlabeled data and further configured to send to a server a parameter of a first subnetwork in the first model; and the server configured to train a second model according to the parameter of the first subnetwork and labeled data to update a parameter of the second model and further configured to send to the client the updated parameter of the first subnetwork and a parameter of a third subnetwork. The client is further configured to obtain a target model according to the parameter of the first subnetwork and the parameter of the third subnetwork from the server. By using the means, the security of private data of the client is guaranteed, moreover, the feature extraction capability of a model on unlabeled data is improved, and labor costs are saved.

Inventors:
ZHANG QI (CN)
WU TIANCHENG (CN)
ZHOU PEICHEN (CN)
Application Number:
PCT/CN2022/095802
Publication Date:
December 22, 2022
Filing Date:
May 28, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
HUAWEI TECH CO LTD (CN)
International Classes:
G06N20/00
Foreign References:
CN112887145A2021-06-01
CN111275207A2020-06-12
US20200218937A12020-07-09
CN202110662048A2021-06-15
Other References:
YUCHEN ZHAO; HANYANG LIU; HONGLIN LI; PAYAM BARNAGHI; HAMED HADDADI: "Semi-supervised Federated Learning for Activity Recognition", ARXIV.ORG, 2 November 2020 (2020-11-02), XP081805371
WONYONG JEONG; JAEHONG YOON; EUNHO YANG; SUNG JU HWANG: "Federated Semi-Supervised Learning with Inter-Client Consistency & Disjoint Learning", ARXIV.ORG, 29 March 2021 (2021-03-29), XP081899292
Attorney, Agent or Firm:
SCIHEAD IP LAW FIRM (CN)
Download PDF: