Improved Differentially Private Euclidean Distance Approximation
Summary: Use Kane–Nelson sparse JL sketches (k=Θ(α^-2 log1/β), s=O(α^-1 log1/β)) combined with Laplace/Gaussian mechanisms to produce unbiased, high-utility differentially private sketches for Euclidean distance. Laplace yields pure DP and lower variance than Gaussian when δ < β^{O(1/α)}, and a private FJLT variant trades speed for variance, resolving an open question of Kenthapadi et al. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 3 of 3 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,446 | PrivBayes: Private Data Release via Bayesian Networks | 2014 | SIGMOD | 0.0001194108 |
| 1,764 | PriView: Practical Differentially Private Release of Marginal Contingency Tables | 2014 | SIGMOD | 0.00010636626 |
| 2,894 | Pan-private Algorithms Via Statistics on Sketches | 2011 | PODS | 7.9474698e-05 |
Previous
Page 1 / 1
Next