Differentially Private Data-Driven Markov Chain Modeling
Alexander Benvenuti, Brandon Fallin, Calvin Hawkins, Brendan Bialy, Miriam Dennis, Warren Dixon, Matthew Hale · 2026
Read Differentially Private Data-Driven Markov Chain Modeling by Alexander Benvenuti, Brandon Fallin, Calvin Hawkins, Brendan Bialy, Miriam Dennis, Warren Dixon, Matthew Hale (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-22443.txt.
Description
Markov chains model user behavior everywhere, but building them needs substantial user data, and sharing them can leak private information. Alexander Benvenuti and six co-authors develop a method for protecting the user data behind a Markov chain model. They start by privatizing database queries whose outputs are probability vectors, proving the method differentially private.
Accuracy is quantified by bounding the expected divergence between private and non-private queries. The method extends to stochastic matrices, whose rows are simplex-valued queries, covering data-driven Markov models. The team further bounds changes in stationary distribution and convergence rate between private and non-private chains. Simulations under typical privacy settings confirm the theory holds in practice.
Catalog record
| CREATOR | Alexander Benvenuti, Brandon Fallin, Calvin Hawkins, Brendan Bialy, Miriam Dennis, Warren Dixon, Matthew Hale |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-22443