Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimizing Privacy and Utility Tradeoffs for Group Interests Through Harmonization

Bishwas Mandal, George Amariucai, Shuangqing Wei

TL;DR

This work introduces a collaborative data-sharing mechanism between two user groups through a trusted third party that uses adversarial privacy techniques with the proposed data-sharing mechanism to internally sanitize data for both groups and eliminates the need for manual annotation or auxiliary datasets.

Abstract

We propose a novel problem formulation to address the privacy-utility tradeoff, specifically when dealing with two distinct user groups characterized by unique sets of private and utility attributes. Unlike previous studies that primarily focus on scenarios where all users share identical private and utility attributes and often rely on auxiliary datasets or manual annotations, we introduce a collaborative data-sharing mechanism between two user groups through a trusted third party. This third party uses adversarial privacy techniques with our proposed data-sharing mechanism to internally sanitize data for both groups and eliminates the need for manual annotation or auxiliary datasets. Our methodology ensures that private attributes cannot be accurately inferred while enabling highly accurate predictions of utility features. Importantly, even if analysts or adversaries possess auxiliary datasets containing raw data, they are unable to accurately deduce private features. Additionally, our data-sharing mechanism is compatible with various existing adversarially trained privacy techniques. We empirically demonstrate the effectiveness of our approach using synthetic and real-world datasets, showcasing its ability to balance the conflicting goals of privacy and utility.

Optimizing Privacy and Utility Tradeoffs for Group Interests Through Harmonization

TL;DR

This work introduces a collaborative data-sharing mechanism between two user groups through a trusted third party that uses adversarial privacy techniques with the proposed data-sharing mechanism to internally sanitize data for both groups and eliminates the need for manual annotation or auxiliary datasets.

Abstract

We propose a novel problem formulation to address the privacy-utility tradeoff, specifically when dealing with two distinct user groups characterized by unique sets of private and utility attributes. Unlike previous studies that primarily focus on scenarios where all users share identical private and utility attributes and often rely on auxiliary datasets or manual annotations, we introduce a collaborative data-sharing mechanism between two user groups through a trusted third party. This third party uses adversarial privacy techniques with our proposed data-sharing mechanism to internally sanitize data for both groups and eliminates the need for manual annotation or auxiliary datasets. Our methodology ensures that private attributes cannot be accurately inferred while enabling highly accurate predictions of utility features. Importantly, even if analysts or adversaries possess auxiliary datasets containing raw data, they are unable to accurately deduce private features. Additionally, our data-sharing mechanism is compatible with various existing adversarially trained privacy techniques. We empirically demonstrate the effectiveness of our approach using synthetic and real-world datasets, showcasing its ability to balance the conflicting goals of privacy and utility.
Paper Structure (20 sections, 3 equations, 17 figures, 3 tables, 2 algorithms)

This paper contains 20 sections, 3 equations, 17 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Formulation
  4. Assumptions and Threat Model
  5. Philosophical Disagreements
  6. Real World Application
  7. Methodology
  8. ALFR and UAE-PUPET
  9. Data Sharing Mechanism
  10. Experimental Setup
  11. Dataset
  12. Private and Utility features
  13. Performance Evaluation Metrics
  14. Result and Discussions
  15. Accuracy and AUROC
  16. ...and 5 more sections

Figures (17)

  • Figure 1: Overview of the privacy-utility tradeoff in two-user group setting. Restricted Access block demonstrates how data from one group trains the privacy mechanism for another and vice-versa. Open Access block details the public release of sanitized data available for public analysis. Blank spaces represent private and utility features of a particular group which are not present in the dataset, and analysts or adversaries aim to make correct predictions of these features.
  • Figure 2: ALFR and UAE-PUPET architecture
  • Figure 3: Sequential and iterative nature of the data sharing approach.
  • Figure 4: US Census - G1
  • Figure 5: US Census - G2
  • ...and 12 more figures