Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Extremal Mechanisms for Local Differential Privacy
Authors: Peter Kairouz, Sewoong Oh, Pramod Viswanath
JMLR 2016 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | For 100 instances of randomly chosen P0 and P1 defined over an input alphabet of size |X| = 6, we compare the average performance of the binary, randomized response, and geometric mechanisms to the average performance of the optimal staircase mechanism for various values of ε. The left panel of Figure 3 shows the average performance measured by the normalized divergence Dkl(M0||M1)/Dkl(P0||P1) for all 4 mechanisms. |
| Researcher Affiliation | Academia | Peter Kairouz EMAIL Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign Urbana, IL 61801, USA Sewoong Oh EMAIL Department of Industrial and Enterprise Systems Engineering University of Illinois at Urbana-Champaign Urbana, IL 61820, USA Pramod Viswanath EMAIL Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign Urbana, IL 61801, USA |
| Pseudocode | No | The paper describes mechanisms (e.g., binary, randomized response, staircase) using mathematical formulations and definitions (e.g., equations for Q(y|x) or matrix representations), but it does not present these in a structured pseudocode or algorithm block. |
| Open Source Code | No | The paper does not contain any statements about releasing code, nor does it provide links to source code repositories or supplementary materials containing code. |
| Open Datasets | No | The paper describes generating '100 instances of randomly chosen P0 and P1' for its numerical experiments, implying data is generated for the study rather than using a pre-existing publicly available dataset with concrete access information. |
| Dataset Splits | No | The paper's 'Numerical Experiments' section describes generating '100 instances of randomly chosen P0 and P1' for comparison. It does not involve traditional machine learning datasets that would typically require train/test/validation splits. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory specifications) used to run the numerical experiments. |
| Software Dependencies | No | The paper does not mention any specific software or library names with version numbers used for its numerical experiments or analysis. |
| Experiment Setup | Yes | For 100 instances of randomly chosen P0 and P1 defined over an input alphabet of size |X| = 6, we compare the average performance of the binary, randomized response, and geometric mechanisms to the average performance of the optimal staircase mechanism for various values of ε. ... For each instance of the 100 randomly generated P0 and P1 defined over input alphabets of size k = 6, we plot the resulting divergence Dkl(M0||M1) as a function of P0 P1 TV for ε = 0.1, and as a function of Dkl(P0||P1) for ε = 10. ... For various input alphabet size |X| {3, 4, 6, 12}, we plot the performance of this mixed strategy for each value of ε and each of the 100 randomly generated instances of P0 and P1. |