Rate Optimality and Phase Transition for User-Level Local Differential Privacy

13.4 MB of research materials on user-level local differential privacy, authored by Alexander Kent and last updated in June 2026. The work derives minimax lower bounds and matching upper bounds for estimation problems, including mean and density estimation, under privacy constraints. It demonstrates phase transition phenomena as the number of observations per user varies.

Use Cases

• Benchmarking privacy-preserving algorithms based on the derived minimax lower bounds for user-level estimation.
• Studying phase transition behavior in estimation risk based on the number of observations per user (T) described in the paper.
• Investigating the feasibility of high-dimensional sparse mean estimation under local privacy constraints, as demonstrated in the work.

Strengths

• Theoretical results include matching lower and upper bounds for several estimation problems, providing a strong analytical foundation.
• The 13.4 MB package includes PDF, ZIP, and TXT files, likely containing the full paper and supplementary materials.
• Released under a permissive CC-BY-4.0 license, allowing for broad reuse and distribution.

Limitations

• The dataset consists of research documents; the underlying numerical or tabular data used for the theoretical analysis is not provided.
• Row count and column-level documentation are unknown, limiting direct analysis of empirical data.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare

Collection Method

Likely contains the authored research paper and supplementary materials.

Freshness

Last updated 2026-06-02 15:18:00; freshness should be verified.