Search papers, labs, and topics across Lattice.
This paper introduces a differentially private method for constructing Markov chain models from user data, addressing the privacy risks associated with sharing such models. They develop a mechanism for privatizing simplex-valued database queries and extend it to privatize stochastic matrices representing Markov chains, providing differential privacy guarantees. The authors analytically bound the KL divergence between private and non-private queries, as well as the changes in stationary distribution and convergence rate, demonstrating empirically that their method achieves high accuracy with minimal impact on the model's behavior.
Achieve differentially private Markov chain modeling with minimal impact on accuracy, showing less than 2% error in the stationary distribution under typical privacy settings.
Markov chains model a wide range of user behaviors. However, generating accurate Markov chain models requires substantial user data, and sharing these models without privacy protections may reveal sensitive information about the underlying user data. We introduce a method for protecting user data used to formulate a Markov chain model. First, we develop a method for privatizing database queries whose outputs are elements of the unit simplex, and we prove that this method is differentially private. We quantify its accuracy by bounding the expected KL divergence between private and non-private queries. We extend this method to privatize stochastic matrices whose rows are each a simplex-valued query of a database, which includes data-driven Markov chain models. To assess their accuracy, we analytically bound the change in the stationary distribution and the change in the convergence rate between a non-private Markov chain model and its private form. Simulations show that under a typical privacy implementation, our method yields less than 2% error in the stationary distribution, indicating that our approach to private modeling faithfully captures the behavior of the systems we study.
