Incentive method and system for hierarchical federated learning under end-side cloud architecture and complete information

A cloud architecture and complete technology, applied in the field of incentive methods and systems of hierarchical federated learning under the cloud architecture and complete information, can solve the problem that the HFL system cannot make good use of the terminal equipment to enrich the data, and reduce the operation time. Time and number of local iterations, cost savings, data privacy protection

Active Publication Date: 2022-01-28
TIANJIN UNIV
View PDF7 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the existing HFL system cannot make good use of the rich data of end-side devices, the present invention proposes a device-edge-cloud architecture and an incentive method and system for hierarchical federated learning under complete information

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
  • Incentive method and system for hierarchical federated learning under end-side cloud architecture and complete information
  • Incentive method and system for hierarchical federated learning under end-side cloud architecture and complete information
  • Incentive method and system for hierarchical federated learning under end-side cloud architecture and complete information

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0071] Embodiment 1: A device-edge-cloud architecture and an incentive method for hierarchical federated learning under complete information, including the following steps:

[0072] S1, based on the Stackelberg game to establish a three-tier game model between terminal devices, edge aggregators and cloud servers;

[0073] Such as figure 1 As shown, the cloud server, edge aggregator and terminal device constitute a hierarchical federated learning system, and the cloud server is the leader, task initiator, and model owner. The edge aggregator runs between the cloud server and multiple terminal devices as a medium for parameter aggregation. The collection of all terminal devices participating in model training adopts Represents, and each terminal device n maintains a data set χ n , x n is the data set χ contributed by terminal device n participating in model training n The amount of data contribution, that is, the strategy of terminal device n. if x n =0, indicating that ...

Embodiment 2

[0161] Embodiment 2: A device-edge-cloud architecture and an incentive system for hierarchical federated learning under complete information, including a cloud server connected to several edge aggregators, and each edge aggregator is connected to several terminal devices , and the cloud server is equipped with a power data learning model based on a federated learning algorithm. The edge aggregator is provided with a first revenue distribution module for distributing remuneration to the connected terminal equipment and a strategy formulation module for evaluating the data contribution of the terminal equipment, and the cloud server is provided with a module for sending the edge aggregator The second revenue distribution module for distributing rewards. During terminal device training, the first revenue distribution module distributes economic benefits to corresponding terminal devices according to the data contribution amount of the terminal device and the data quality of the co...

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

The invention discloses an incentive method and system for hierarchical federated learning under an end-side cloud architecture and complete information. The method comprises the following steps: establishing a three-layer game model among terminal equipment, an edge aggregator and a cloud server based on a Stackelberg game; enabling the terminal equipment to download the power data learning model in the cloud server through the edge aggregator; constructing a model evolution function based on a hierarchical federated learning algorithm; solving an optimal strategy of three parties by utilizing Nash equilibrium, so the utility of the terminal equipment, the utility of the edge aggregator and the utility of the cloud server are maximized; enabling the terminal terminal to send the model updated based on the optimal strategy to the edge aggregator, enabling the edge aggregator to carry out excitation issuing on the terminal device based on the optimal strategy and sending the aggregated model to the cloud server; and enabling the cloud server to carry out excitation issuing on the edge aggregator based on the optimal strategy and carry out re-aggregation on the model. According to the invention, rewards can be provided for the terminal equipment and the edge aggregator, and reasonable and fair income distribution is realized.

Description

technical field [0001] The invention belongs to the technical field of federated learning, and in particular relates to a device-edge-cloud architecture and an incentive method and system for layered federated learning under complete information. Background technique [0002] Recent developments in deep learning have revolutionized many application domains, such as image processing, natural language processing, video analysis, etc., including the field of electricity. The great success of deep learning in these fields stems from the availability of large amounts of training data and massive computing power. However, training data is generated by distributed devices owned by individuals or different organizations. If such data is leaked or used for other purposes than the original purpose, personal privacy will be compromised. For example, some power data involves personal user privacy information and has high security requirements. Once the privacy, integrity, and availabi...

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 Applications(China)
IPC IPC(8): H04L67/10H04L67/1004G06N20/00G06N3/04G06N3/08
CPCH04L67/10H04L67/1004G06N20/00G06N3/08G06N3/045Y04S10/50
Inventor 王晓飞赵云凤刘志成仇超
Owner TIANJIN UNIV
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