User behavior recognition method based on personalized semi-supervised online federated learning

A recognition method, semi-supervised technology, applied in neural learning methods, character and pattern recognition, climate sustainability, etc., can solve problems such as scarcity of privacy-preserving labels and real-time, and achieve the goal of overcoming convergence instability and conceptual drift. Effect

Active Publication Date: 2021-08-24
SHANDONG UNIV
View PDF7 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Especially to make user behavior recognition methods applicable to real-world scenarios, there are four challenges to be addressed, namely, privacy protection, label scarcity, real-time and heterogeneity

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
  • User behavior recognition method based on personalized semi-supervised online federated learning
  • User behavior recognition method based on personalized semi-supervised online federated learning
  • User behavior recognition method based on personalized semi-supervised online federated learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] Embodiment 1 This embodiment provides a semi-supervised online learning-based personalized federated user behavior recognition method, the process is as follows figure 1 As shown, it mainly includes the following steps:

[0041] One: Determine the tagged client and the untagged client, and make preparations.

[0042] In this step, it is first necessary to determine the information of each client in the federated learning. In this embodiment, clients are divided into two categories, one is a client with a dataset of labeled samples (hereinafter referred to as a labeled client), and the other is a client with an unlabeled data stream ( Hereinafter referred to as unlabeled client). In any labeled client, there is a local dataset where the samples in the data are labeled. For any unlabeled client, there is a data stream, and the data in the data stream is unlabeled. This is actually very much in line with real-life scenarios, where smart devices worn by people generate ...

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 belongs to the technical field of intelligent equipment user behavior recognition, and particularly relates to a user behavior recognition method based on personalized semi-supervised online federated learning. The method is characterized by comprising the following steps: determining a client with a label and a client without a label; performing online semi-supervised federated learning by a FedHAR algorithm, and training a generalized neural network model; and performing personalized federal fine tuning on the generalized neural network model to obtain a multi-modal personalized neural network model. A semi-supervised online learning and federated learning framework is used, and a personalized user behavior recognition method is provided to solve the behavior recognition problem in a real scene and the privacy problem.

Description

technical field [0001] The invention belongs to the technical field of user behavior recognition of smart devices, and in particular relates to a user behavior recognition method based on personalized semi-supervised online federated learning. Background technique [0002] With the development of sensor technology and the improvement of mobile phone computing power, user behavior recognition based on smartphone sensors has become a research hotspot in recent years. User behavior recognition takes the raw data of mobile phone or wearable device sensor as input, and predicts the user's motion behavior through the recognition algorithm. It plays an important role in applications such as health and activity monitoring, user biometric signatures, urban computing, assistance for the disabled, elderly care, and indoor positioning. At present, most sensor-based user behavior recognition research focuses on the recognition of simple behaviors. In many studies, the accuracy rate of s...

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): G06K9/62G06N3/08
CPCG06N3/08G06F18/2155Y02D10/00
Inventor 张啸于宏正
Owner SHANDONG 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