Unlock instant, AI-driven research and patent intelligence for your innovation.

Efficient data perception layered federated learning method based on task unloading

A technology of data perception and learning methods, applied in the field of joint task offloading, which can solve problems such as expensive, serious, and limited communication resources

Pending Publication Date: 2022-07-29
SHANGHAI TECH UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the communication resources in the wide area network (hereinafter referred to as WAN) in the two-layer C-FL framework are limited and expensive
Network congestion is exacerbated when a large number of devices communicate with the cloud through the backbone network

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
  • Efficient data perception layered federated learning method based on task unloading
  • Efficient data perception layered federated learning method based on task unloading
  • Efficient data perception layered federated learning method based on task unloading

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0085] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these examples are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of the present application.

[0086] 1 System Model

[0087] 1.1 Application scenarios

[0088] FL is an exploration of distributed machine learning that can be trained using scattered data. This concept is well adapted to the characteristics of data sharding in MEC. Therefore, the introduction of FL into MEC has great engineering practical value. figure 1 The offloading of HFEL (hereinafter referred to as TO) and bandwidth allocation (hereinafter referr...

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 technical scheme of the invention provides an efficient data perception hierarchical federal learning method based on task unloading. According to the method, data distribution in the cost function is considered for the first time, and the quality of the edge data set can be improved while the system cost is reduced. In addition, the invention designs a TO and RA method based on a multi-intelligent body depth deterministic strategy gradient model capable of reducing action space. A large number of experiments prove that the algorithm provided by the invention can effectively improve the accuracy of the aggregation model, effectively reduce the unloading cost, improve the training precision of the lightweight data sensing HFEL algorithm and reduce the system cost.

Description

technical field [0001] The present invention relates to joint task offloading, resource allocation and participant selection problems under Hierarchical Federated Edge Learning (hereinafter referred to as HFEL) to reduce system cost and improve the training accuracy of Federated Learning (hereinafter referred to as FL). Background technique [0002] In the era of data intelligence, billions of devices generate massive amounts of data at the edge. Uploading personal data to a third-party cloud server for computing can cause many problems, including privacy breaches. As an effective coping method, FL has emerged as a promising machine learning paradigm. FL aggregates multiple weights by uploading the trained gradients or weights, and finally obtains a global model. FL has been applied to multi-access edge computing (hereafter referred to as MEC) scenarios for distributed model training to protect data privacy. [0003] Traditional FL is dominated by a two-layer cloud federa...

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
IPC IPC(8): H04W28/08
CPCH04W28/0983H04W28/0925
Inventor 马牧雷吴连涛杨旸
Owner SHANGHAI TECH UNIV