Statistics > Machine Learning
[Submitted on 31 Aug 2023 (this version), latest version 3 Mar 2024 (v2)]
Title:Differentially Private Functional Summaries via the Independent Component Laplace Process
View PDFAbstract:In this work, we propose a new mechanism for releasing differentially private functional summaries called the Independent Component Laplace Process, or ICLP, mechanism. By treating the functional summaries of interest as truly infinite-dimensional objects and perturbing them with the ICLP noise, this new mechanism relaxes assumptions on data trajectories and preserves higher utility compared to classical finite-dimensional subspace embedding approaches in the literature. We establish the feasibility of the proposed mechanism in multiple function spaces. Several statistical estimation problems are considered, and we demonstrate by slightly over-smoothing the summary, the privacy cost will not dominate the statistical error and is asymptotically negligible. Numerical experiments on synthetic and real datasets demonstrate the efficacy of the proposed mechanism.
Submission history
From: Haotian Lin [view email][v1] Thu, 31 Aug 2023 20:24:51 UTC (8,912 KB)
[v2] Sun, 3 Mar 2024 23:49:48 UTC (8,937 KB)
Current browse context:
stat.ML
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.