Differentially Private Explanations for Clusters

TL;DR

This paper introduces DPClustX, a framework that offers privacy-preserving explanations for clustering results under differential privacy, enabling analysts to interpret clusters without compromising sensitive data.

Contribution

The paper presents DPClustX, a novel method for generating global explanations of clustering results that satisfies differential privacy, addressing the challenge of interpretability under privacy constraints.

Findings

DPClustX provides accurate explanations under strict privacy budgets.

The framework effectively balances interpretability and privacy in clustering analysis.

Experimental results demonstrate the utility of DPClustX on real datasets.

Abstract

The dire need to protect sensitive data has led to various flavors of privacy definitions. Among these, Differential privacy (DP) is considered one of the most rigorous and secure notions of privacy, enabling data analysis while preserving the privacy of data contributors. One of the fundamental tasks of data analysis is clustering , which is meant to unravel hidden patterns within complex datasets. However, interpreting clustering results poses significant challenges, and often necessitates an extensive analytical process. Interpreting clustering results under DP is even more challenging, as analysts are provided with noisy responses to queries, and longer, manual exploration sessions require additional noise to meet privacy constraints. While increasing attention has been given to clustering explanation frameworks that aim at assisting analysts by automatically uncovering the characteristics of each cluster, such frameworks may also disclose sensitive information within the dataset, leading to a breach in privacy. To address these challenges, we present DPClustX, a framework that provides explanations for black-box clustering results while satisfying DP. DPClustX takes as input the sensitive dataset alongside privately computed clustering labels, and outputs a global explanation, emphasizing prominent characteristics of each cluster while guaranteeing DP. We perform an extensive experimental analysis of DPClustX on real data, showing that it provides insightful and accurate explanations even under tight privacy constraints.