Privacy-Aware Location Sharing with Deep Reinforcement Learning

Abstract

Internet of things (IoT) devices are becoming increasingly popular thanks to many new services and applications they offer. However, in addition to their many benefits, they raise privacy concerns since they share fine-grained time-series user data with untrusted third parties. In this work, we study the privacy-utility trade-off (PUT) in time-series data sharing. Existing approaches to PUT mainly focus on a single data point; however, temporal correlations in time-series data introduce new challenges. Methods that preserve the privacy for the current time may leak significant amount of information at the trace level as the adversary can exploit temporal correlations in a trace. We consider sharing the distorted version of a user's true data sequence with an untrusted third party. We measure the privacy leakage by the mutual information between the user's true data sequence and shared version. We consider both instantaneous and average distortion between the two sequences, under a given distortion measure, as the utility loss metric. To tackle the historydependent mutual information minimization, we reformulate the problem as a Markov decision process (MDP), and solve it using asynchronous actor-critic deep reinforcement learning (RL). We apply our optimal data release policies to location trace privacy scenario, and evaluate the performance of the proposed policy numerically.

Active Privacy-Utility Trade-off Against Inference in Time-Series Data Sharing

“…Among those that consider temporal correlations, most existing works focus on the privacy of the time-series measurements rather than hiding latent sensitive attributes [9], [10], [26]- [30]. In the location sharing scenario, sensitive information is the time-series data itself and the utility loss can be measured by data distortion, whereas in many applications, the user might be interested in hiding an underlying sensitive hypothesis.…”

“…Differential privacy (DP), k-anonymity, information theoretic metrics and the SP's error probability are commonly used as privacy measures [8], [9], [11], [15], [20]- [31], [38]. By definition, DP prevents the SP from inferring the current data of the user, even if the SP has the knowledge of all the remaining data points.…”

“…ITP and utility for location sharing is studied in [9], and extended to generic time-series data release in [10]. The user follows a history-dependent online data release policy by minimizing the MI between the real and modified location trajectories subject to a distortion constraint.…”

Active Privacy-Utility Trade-Off Against A Hypothesis Testing Adversary

“…Privacy is an important concern for the adoption of many IoT services, and there is a growing demand from consumers to keep their personal information private. Privacy has been widely studied in the literature [1][2][3][4][5][6][7][8][9][10], and a vast number of privacy measures have been introduced, including differential privacy [1], mutual information (MI) [2][3][4][5][6][7][8], total variation distance [11], maximal leakage [12,13], and guessing leakage [14], to count a few.…”

Active Privacy-utility Trade-off Against a Hypothesis Testing Adversary

“…Privacy is an important concern for the adoption of many IoT services, and there is a growing demand from consumers to keep their personal information private. Privacy has been widely studied in the literature [1][2][3][4][5][6][7][8][9][10], and a vast number of privacy measures have been introduced, including differential privacy [1], mutual information (MI) [2][3][4][5][6][7][8], total variation distance [11], maximal leakage [12,13], and guessing leakage [14], to count a few.…”