Abstract :

With Innovation Based On Available Data, The Creation Of Machine Learning Systems That Ensures User Privacy Is Now More Pressing Than Ever. Organizations Have Been Utilizing Centralized Datasets For Model Training Purposes Which Raises Serious Concerns About Potential Data Leakage Breaches, Misuse As Well As Regulatory Compliance. The Project “Secure Privacy Using Federated Learning Techniques” Attempts To Solve These Issues By Employing An Approach Based On Decentralization And Privacy-preserving Techniques. Providing Critical Defense Mechanisms On Sensitive Information At The Data And Model Level While Ensuring Performance And Scalability Are Preserved Is The Objective Of This Model. This Particular Project Uses The Paradigm Of Federated Learning (FL), Which Is A Form Of Collaborative Machine Learning Where Model Training Happens On Various Devices Or Servers In A Decentralized Manner, And Raw Data Stays Local. FL Helps Reduce Privacy Risk As It Eliminates The Need To Transfer Private Data To Central Repositories. To Make This Approach Stronger, The System Incorporates Advanced Cryptographic Methods Such As Homomorphic Encryption, Differential Privacy, And Secure Multi-Party Computation (SMPC). These Approaches Provide Encrypted Calculations, Data Mask Through Noise Addition As Well As Secure Collective Computations Thereby Protecting Each Individual Data Point From Exposure During Training Agnostic Multidimensional Space Process Known As Computation Graph Directed Acyclic Graphs,. Besides Privacy Preservation, Model Failure Reasons Are Also Addressed By This Project Key Words: Secure Multi-Party Computation, Differential Privacy, Homomorphic Encryption, Federated Learning.

Published:

11-7-2025

Issue:

Vol. 25 No. 7 (2025)

Page Nos:

262-271

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite