Title: Machine Learning Approaches That Keep User Data Decentralized To Enhance Privacy And Comply With Regulations Like Gdpr
Authors: Berdiyev Usmon Tolib o'g'li and Allanazarov Ravshan Shavkat o'g'li
Volume: 9
Issue: 8
Pages: 45-49
Publication Date: 2025/08/28
Abstract:
With the exponential growth of data-driven applications, privacy concerns have become a central issue in machine learning (ML). Regulations such as the General Data Protection Regulation (GDPR) have further emphasized the need for systems that protect individual data. This paper explores decentralized machine learning methods that enhance user data privacy and ensure regulatory compliance. Techniques such as federated learning, split learning, and homomorphic encryption are examined in terms of their privacy guarantees, efficiency, and practical deployment. The discussion also evaluates trade-offs and challenges associated with decentralization, such as communication overhead and data heterogeneity.