Research map / Federated learning and privacy / Differential privacy
Differential Privacy Over Riemannian Manifolds
Matthew Reimherr, Karthik Bharath, Carlos J Soto
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| privatize via intrinsic Riemannian Laplace instead of ambient Euclidean Laplace | Frechet mean privatization · SPDM P(k) | Euclidean distance between mean and private mean | Improved |
| privatize via intrinsic Riemannian Laplace instead of ambient Euclidean Laplace | Frechet mean privatization · Sphere S^2 | average Euclidean distance (noise): 16.8% | Improved |
| privatize via intrinsic Riemannian Laplace instead of ambient Euclidean Laplace at larger sample sizes | Frechet mean privatization · Sphere S^2 | average Euclidean distance (noise): 12% | Improved |
| use ambient Euclidean embedding instead of intrinsic mechanism | SPDM private mean · SPDM P(2), sample sizes 20-40 | fraction valid on manifold: 25% | Hurt |
Closest papers by meaning
- Shape And Structure Preserving Differential Privacy (NeurIPS 2022 · similarity 0.93)
- Gaussian Differential Privacy on Riemannian Manifolds (NeurIPS 2023 · similarity 0.93)
- Exponential-Wrapped Mechanisms: Differential Privacy on Hadamard Manifolds Made Practical (ICLR 2026 · similarity 0.91)
- Wrapped Gaussian on the manifold of Symmetric Positive Definite Matrices (ICML 2025 · similarity 0.86)
- On Differentially Private Subspace Estimation in a Distribution-Free Setting (NeurIPS 2024 · similarity 0.85)
- Intrinsic Riemannian Cross-covariance for Manifold-valued Random Objects (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
