On Differentially Private U Statistics
Authors: Kamalika Chaudhuri, Po-Ling Loh, Shourya Pandey, Purnamrita Sarkar
NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | We consider the problem of privately estimating a parameter E[h(X1, . . . , Xk)]... To remedy this, we propose a new thresholding-based approach... This leads to nearly optimal private error for non-degenerate U-statistics and a strong indication of near-optimality for degenerate U-statistics. Our contributions 1. We present a new algorithm for private mean estimation... 2. We provide a lower bound... 3. The computational complexity... |
| Researcher Affiliation | Academia | Kamalika Chaudhuri Po-Ling Loh University of California San Diego University of Cambridge kamalika@ucsd.edu pll28@cam.ac.uk Shourya Pandey Purnamrita Sarkar University of Texas at Austin University of Texas at Austin shouryap@utexas.edu purna.sarkar@austin.utexas.edu |
| Pseudocode | Yes | Algorithm 1 Private Mean Local Hájek(n, k, {h(XS), S S}, ϵ, C, ξ, S) ... Algorithm A.2 U-Stat Mean n, k, h, {Xi}i [n] , F = {S1, . . . , Sm}, R, ϵ, γ, Q( ), Qavg( ) ... Algorithm A.3 U-Stat One Step n, k, {Yi}i [m] , F, [l, r], ϵ , β, Q( ), Qavg( ) |
| Open Source Code | No | The paper does not provide an explicit statement of code release, a link to a code repository, or mention of code in supplementary material. The NeurIPS checklist also indicates 'NA' for open access to code. |
| Open Datasets | No | The paper is theoretical and focuses on algorithm design and mathematical proofs, not on empirical studies with datasets. Thus, it does not provide information about public datasets used for training. The NeurIPS checklist indicates 'NA' for all experimental questions. |
| Dataset Splits | No | The paper is theoretical and does not report on empirical experiments with data splits. The NeurIPS checklist indicates 'NA' for experimental questions, confirming no such splits are provided. |
| Hardware Specification | No | The paper is theoretical and does not describe empirical experiments requiring specific hardware for execution. The NeurIPS checklist indicates 'NA' for experimental questions, including compute resources. |
| Software Dependencies | No | The paper is theoretical and does not report on empirical experiments, therefore it does not list specific software dependencies with version numbers required for reproducibility. The NeurIPS checklist indicates 'NA' for all experimental questions. |
| Experiment Setup | No | The paper is theoretical and focuses on algorithm design and proofs, not on empirical experiments. Thus, it does not provide details on experimental setup such as hyperparameters or training configurations. The NeurIPS checklist indicates 'NA' for all experimental questions. |