Differential Privacy for Clustering Under Continual Observation
Data Structures and Algorithms
2023-07-28 v2 Cryptography and Security
Machine Learning
Abstract
We consider the problem of clustering privately a dataset in that undergoes both insertion and deletion of points. Specifically, we give an -differentially private clustering mechanism for the -means objective under continual observation. This is the first approximation algorithm for that problem with an additive error that depends only logarithmically in the number of updates. The multiplicative error is almost the same as non privately. To do so we show how to perform dimension reduction under continual observation and combine it with a differentially private greedy approximation algorithm for -means. We also partially extend our results to the -median problem.
Cite
@article{arxiv.2307.03430,
title = {Differential Privacy for Clustering Under Continual Observation},
author = {Max Dupré la Tour and Monika Henzinger and David Saulpic},
journal= {arXiv preprint arXiv:2307.03430},
year = {2023}
}