Analysis, Design, and Implementation of a User-Friendly Differential Privacy Application DOI Creative Commons

Reynardo Tjhin,

Muhammad Sajjad Akbar, Clément L. Canonne

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(5), P. 1358 - 1358

Published: Feb. 23, 2025

In the era of artificial intelligence, ensuring privacy in publicly released data is critical to prevent linkage attacks that can reveal sensitive information about individuals. Differential (DP) has emerged as a robust approach for safeguarding privacy, but its mathematical complexity often limits accessibility non-experts. This paper introduces novel, user-friendly web application bridges gap between theoretical DP concepts and their practical application. The includes two main features: query version, which demonstrates mechanisms statistical queries; privatize applies techniques entire datasets. A key contribution this work identification discrepancies implementation maximum minimum queries within OpenDP library, revealing gaps theory practice. Additionally, foundational framework dataset privatization using OpenDP’s built-in methods. By providing an interactive platform, advances public understanding highlights areas improvement existing libraries. serves both educational tool step toward addressing challenges DP.

Language: Английский

Analysis, Design, and Implementation of a User-Friendly Differential Privacy Application DOI Creative Commons

Reynardo Tjhin,

Muhammad Sajjad Akbar, Clément L. Canonne

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(5), P. 1358 - 1358

Published: Feb. 23, 2025

In the era of artificial intelligence, ensuring privacy in publicly released data is critical to prevent linkage attacks that can reveal sensitive information about individuals. Differential (DP) has emerged as a robust approach for safeguarding privacy, but its mathematical complexity often limits accessibility non-experts. This paper introduces novel, user-friendly web application bridges gap between theoretical DP concepts and their practical application. The includes two main features: query version, which demonstrates mechanisms statistical queries; privatize applies techniques entire datasets. A key contribution this work identification discrepancies implementation maximum minimum queries within OpenDP library, revealing gaps theory practice. Additionally, foundational framework dataset privatization using OpenDP’s built-in methods. By providing an interactive platform, advances public understanding highlights areas improvement existing libraries. serves both educational tool step toward addressing challenges DP.

Language: Английский

Citations

0