Sublinear Space Private Algorithms Under the Sliding Window Model
[edit]
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:63636372, 2019.
Abstract
The Differential privacy overview of Apple states, “Apple retains the collected data for a maximum of three months." Analysis of recent data is formalized by the sliding window model. This begs the question: what is the price of privacy in the sliding window model? In this paper, we study heavy hitters in the sliding window model with window size $w$. Previous works of Chan et al. (2012) estimates heavy hitters with an error of order $\theta w$ for a constant $\theta >0$. In this paper, we give an efficient differentially private algorithm to estimate heavy hitters in the sliding window model with $\widetilde O(w^{3/4})$ additive error and using $\widetilde O(\sqrt{w})$ space.
Related Material


