Interpretable Machine Learning for Privacy-Preserving Pervasive Systems

TL;DR

This paper introduces an interpretability framework for machine learning in pervasive systems, helping users understand how their behavioral traces may compromise their privacy.

Contribution

It presents a novel interpretability approach specifically designed for privacy analysis in pervasive machine learning applications.

Findings

Framework effectively reveals privacy violations

Enhances user understanding of trace-based privacy risks

Supports privacy-preserving system design

Abstract

Our everyday interactions with pervasive systems generate traces that capture various aspects of human behavior and enable machine learning algorithms to extract latent information about users. In this paper, we propose a machine learning interpretability framework that enables users to understand how these generated traces violate their privacy.