KNG: The K-Norm Gradient Mechanism

Authors: Matthew Reimherr, Jordan Awan

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In addition to theoretical guarantees on privacy and utility, we confirm the utility of KNG empirically in the settings of linear and quantile regression through simulations.
Researcher Affiliation Academia Matthew Reimherr Department of Statistics Pennsylvania State University State College, PA 16802 mreimherr@psu.edu Jordan Awan Department of Statistics Pennsylvania State University State College, PA 16802 awan@psu.edu
Pseudocode Yes Algorithm 1 Regression Simulation
Open Source Code No The paper discusses the implementation of sampling procedures (e.g., MCMC) but does not provide any link or explicit statement about releasing the source code for the methodology described.
Open Datasets No The paper describes generating synthetic data for simulations using specific distributions (e.g., Xij iid U(-1, 1), errors ei N(0, 1)) rather than using a publicly available or open dataset.
Dataset Splits No The paper conducts simulations by generating data for each replicate but does not describe the use of explicit training, validation, and test dataset splits from a pre-existing dataset.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments.
Software Dependencies No The paper mentions using a 'one-at-a-time MCMC procedure' but does not specify any software names with version numbers (e.g., Python, PyTorch, TensorFlow, or specific statistical packages) used for the implementation.
Experiment Setup Yes For each n in 10^2, 10^3, 10^4, . . . , 10^7 we run 100 replicates of Algorithm 1 at ϵ = 1. For KNG and exponential mechanism, we draw samples using a one-at-a-time MCMC procedure with 10000 steps.