Locally Differentially-Private Randomized Response for Discrete Distribution Learning
Authors: Adriano Pastore, Michael Gastpar
JMLR 2021 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | We derive the respective normalized first-order terms of convergence (as n ), which for a given target privacy ϵ represent a rule-of-thumb factor by which the sample size must be augmented so as to achieve the same estimation accuracy as that of a non-randomizing channel. We formulate the privacy fidelity trade-off problem as being that of minimizing said first-order term under a privacy constraint ϵ. We further identify a scalar quantity that captures the essence of this trade-off, and prove bounds and data-processing inequalities on this quantity. For some specific instances of the privacy fidelity trade-offproblem, we derive inner and outer bounds on the optimal trade-offcurve. |
| Researcher Affiliation | Academia | Adriano Pastore EMAIL Department of Statistical Inference for Communications and Positioning Centre Tecnol ogic de Telecomunicacions de Catalunya (CTTC/CERCA) Castelldefels, Barcelona, Spain Michael Gastpar EMAIL School for Computer and Communication Sciences Ecole Polytechnique F ed erale de Lausanne Lausanne, Switzerland |
| Pseudocode | No | The paper describes methods mathematically and textually, but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code, nor does it provide links to any code repositories for the methodology described. |
| Open Datasets | Yes | in (b) for the age pyramid of the Brazilian population, drawn from the IPUMS data sets (Minnesota Population Center, 2019), similar to the data set used by Wang et al. (2019). |
| Dataset Splits | No | The paper is theoretical and illustrates concepts with numerical evaluations; it does not describe machine learning experiments that would typically require training/test/validation dataset splits. The data mentioned (IPUMS) is used for illustrative purposes with different binning strategies, not for model training/evaluation splits. |
| Hardware Specification | No | The paper describes theoretical derivations and numerical evaluations, but does not specify any particular hardware used for computations or experiments. |
| Software Dependencies | No | The paper does not mention any specific software dependencies with version numbers for reproducing the numerical evaluations or theoretical derivations. |
| Experiment Setup | No | The paper is primarily theoretical, focusing on mathematical derivations and bounds. It includes numerical evaluations of these derivations but does not describe any experimental setup with hyperparameters or training configurations typically found in empirical studies. |