Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Federal learning-based human activity identification method and system

A human activity and recognition method technology, applied in the field of human activity recognition, can solve the problems of inability to protect user data privacy and insufficient recognition result accuracy, and achieve the effect of optimizing model aggregation effect and improving recognition accuracy.

Pending Publication Date: 2022-04-22
HENAN UNIVERSITY
View PDF0 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that the current human activity recognition method cannot protect the privacy of user data or the accuracy of the recognition result is not enough, the present invention provides a human activity recognition method and system based on federated learning, which can improve the accuracy of human recognition while protecting user privacy. Accuracy

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
  • Federal learning-based human activity identification method and system
  • Federal learning-based human activity identification method and system
  • Federal learning-based human activity identification method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 2

[0044] On the basis of the above-mentioned embodiments, the embodiment of the present invention also provides a human activity recognition method based on federated learning (called P-FEDAVG), including the following steps:

[0045] S201: The server determines the training task and the corresponding activity data features, generates initial model parameters, and sends the initial model parameters to each client;

[0046] S202: Each client performs local training according to the received initial model parameters or new global model parameters in combination with its own local activity data, generates local update parameters, and uploads the local update parameters to the server;

[0047] When training in combination with local activity data and current model parameters, the multi-category cross-entropy loss function categorical_crossentropy is used and softmax is introduced as a classifier to identify activities.

[0048] As an implementable mode, each client uses the gradient d...

Embodiment 3

[0065] combine figure 1 As shown, the embodiment of the present invention also provides a human activity recognition system based on federated learning, including: a server and multiple clients; multiple clients establish connections and communicate with the server through Ethernet;

[0066] The server is configured to determine the training tasks and corresponding activity data features, and generate initial model parameters, and send the initial model parameters to each client; and perform weighted average aggregation of the received local update parameters of each client to generate new global model parameters, and send the new global model parameters to each client;

[0067] Each client is configured to perform local training in combination with its own local activity data according to the received initial model parameters or new global model parameters, generate local update parameters, and upload the local update parameters to the server.

[0068] It should be noted tha...

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 provides a human activity identification method and system based on federal learning. The method comprises the steps that 1, a server determines training tasks and corresponding activity data features, generates initial model parameters and issues the initial model parameters to all clients; 2, each client performs local training according to the received initial model parameters or new global model parameters in combination with local activity data of the client, generates local update parameters and uploads the local update parameters to a server; step 3, the server performs weighted average aggregation on the received local update parameters of each client to generate new global model parameters, and issues the new global model parameters to each client; and step 4, repeatedly executing the step 2 to the step 3 until a training stop condition is met to obtain a final global model parameter, and issuing the final global model parameter to each client by the server. According to the method, the accuracy of HAR recognition is improved while the privacy of the user is protected.

Description

technical field [0001] The present invention relates to the technical field of human activity recognition, in particular to a method and system for human activity recognition based on federated learning. Background technique [0002] With the rapid increase of population aging, the research of human activity recognition (HAR) has become particularly important. People can detect daily activities through wearable devices such as smartphones, wristbands and smart glasses. In recent years, machine learning has been increasingly used in the field of human activity recognition. Various methods based on heuristic manual feature extraction and automatic feature extraction have been applied in human activity recognition for recognition classification. Existing human activity recognition work currently focuses on centralized learning methods that train models on a central server, which needs to collect all activity data from users for centralized training. However, the centralized c...

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): G06N20/20G06N3/08G06F21/62
CPCG06N20/20G06N3/08G06F21/6245
Inventor 刘颜红徐恕贞何欣王光辉田贺文
Owner HENAN UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com