Privacy federation learning framework based on principal component analysis

Jiaheng Yang; Xia Feng; Yueming Liu

doi:10.1117/12.3031919

6 June 2024 Privacy federation learning framework based on principal component analysis

Jiaheng Yang, Xia Feng, Yueming Liu

Proceedings Volume 13175, International Conference on Computer Network Security and Software Engineering (CNSSE 2024); 131750E (2024) https://doi.org/10.1117/12.3031919
Event: 4th International Conference on Computer Network Security and Software Engineering (CNSSE 2024), 2024, Sanya, China

Abstract

Federal learning is an effective distributed learning technology that allows machine learning model training while protecting data privacy. However, with the increase of the number of user -side devices, the calculation burden of users in federal learning will increase. Researchers explore the use of dimension reduction technology to reduce the calculation burden required for model training, but this triggers a problem with low accuracy. This article extracts the dimensions of gradient data by improving the main component analysis method to extract the dimensions of gradient data and reduce communication and calculation burden while protecting the privacy of the client. The experimental results of this article show that under large -scale data sets, the method of this article increases the speed of 50%training and reaches 96% accuracy.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Jiaheng Yang, Xia Feng, and Yueming Liu "Privacy federation learning framework based on principal component analysis", Proc. SPIE 13175, International Conference on Computer Network Security and Software Engineering (CNSSE 2024), 131750E (6 June 2024); https://doi.org/10.1117/12.3031919

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available