Differentially Private and Federated Structure Learning in Bayesian Networks
Ghita Fassy El Fehri, Aurélien Bellet, Philippe Bastien · 2025
Read Differentially Private and Federated Structure Learning in Bayesian Networks by Ghita Fassy El Fehri, Aurélien Bellet, Philippe Bastien (2025) 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-2512-01708.txt.
Description
Ghita Fassy El Fehri, Aurelien Bellet, and Philippe Bastien learn Bayesian network structure with privacy and federation. Data sits split across sites that cannot pool it. Privacy rules bind each site. Communication costs also grow fast with dimension.
Their Fed-Sparse-BNSL method targets linear Gaussian networks. It keeps rigorous privacy guarantees for each party. Sparse structure keeps messages small as dimension grows. Joint learning without raw data sharing serves hospitals and banks that must map links in data they cannot merge. Sparse structure priors do double duty, cutting both message size and the risk of spurious links in small local data sets.
Catalog record
| CREATOR | Ghita Fassy El Fehri, Aurélien Bellet, Philippe Bastien |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-01708