A model training method based on federated learning

A model training and federation technology, applied in the information field, can solve the problem that the model parameters are not suitable for exposed nodes and so on

Active Publication Date: 2022-03-25
ALIPAY (HANGZHOU) INFORMATION TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, in some scenarios, model parameters are not suitable for exposure to nodes

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A model training method based on federated learning
  • A model training method based on federated learning
  • A model training method based on federated learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] Under the federated learning framework, the server is usually responsible for updating the model parameters according to the gradient uploaded by the node, and sending the model parameters to the node, and the node calculates the gradient based on the model parameters and local training samples. In order to prevent the server from inferring the local training samples of the node based on the gradient uploaded by the node, the node uploads the gradient to the server based on the SA protocol, so that the server only obtains the sum of the gradients uploaded by each node, but cannot obtain Gradients uploaded by a single node.

[0029] It can be seen that under the existing federated learning architecture, nodes can hide local training samples from the server, but the server will not hide model parameters from the nodes.

[0030] However, in some scenarios, the server does not want to expose privacy (ie, model parameters) to nodes. For example, suppose it is necessary to t...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

A model training method based on federated learning is disclosed. In order to protect the privacy of the server (model parameters) from being leaked, the server uses a homomorphic encryption algorithm to encrypt the model parameter set and sends it to the node. Based on the principle of homomorphic encryption, the node uses the encrypted model parameters and local training samples to encrypt The model calculation in the state obtains the encrypted gradient. Then, based on the principle of homomorphic encryption, the node calculates the difference between the encrypted gradient and the encrypted random number, which is essentially a meaningless encrypted value. Then, the node uploads the encrypted value to the server. In addition, the server can use the SA protocol to obtain the sum of random numbers on each node without knowing the random number on each node. In this way, the server can restore the sum of the gradients generated by each node based on the encrypted value uploaded by each node and the sum of each random number, so that the model parameters can be updated.

Description

technical field [0001] The embodiment of this specification relates to the field of information technology, and in particular to a model training method based on federated learning. Background technique [0002] Federated learning (Federated machine learning / Federated Learning) refers to a machine learning framework that can effectively help multiple nodes (which can represent individuals or institutions) jointly train models while meeting the requirements of data privacy protection. [0003] Under the framework of federated learning, the server sends model parameters to multiple nodes, and each node inputs local training samples into the model for a training session. After the training, each node will calculate the gradient based on the training results . Subsequently, the server can calculate the sum of the gradients of each node based on the Secure Aggregation (SA, Secure Aggregation) protocol. It is worth emphasizing that the server is restricted by the SA protocol and...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F21/60G06F21/62G06N20/00
CPCG06F21/602G06F21/6245G06N20/00
Inventor 王力陈超超周俊
Owner ALIPAY (HANGZHOU) INFORMATION TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap