Private Edge Density Estimation for Random Graphs: Optimal, Efficient and Robust

Authors: Hongjie Chen, Jingqiu Ding, Yiding Hua, David Steurer

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We give the first polynomial-time, differentially node-private, and robust algorithm for estimating the edge density of Erdős-Rényi random graphs and their generalization, inhomogeneous random graphs. We further prove information-theoretical lower bounds, showing that the error rate of our algorithm is optimal up to logarithmic factors. Previous algorithms incur either exponential running time or suboptimal error rates.
Researcher Affiliation Academia Hongjie Chen ETH Zürich Jingqiu Ding ETH Zürich Yiding Hua ETH Zürich David Steurer ETH Zürich
Pseudocode Yes Algorithm 2.5 (Private coarse estimation for Erdős-Rényi random graphs). Input: 𝜂-corrupted Erdős-Rényi random graph 𝐴. Privacy parameter: 𝜀. Output: A sample from the distribution 𝜇𝐴,𝜀with support [0, 𝑛] and density d𝜇𝐴,𝜀(𝑑) exp( 𝜀 sos-score(𝑑; 𝐴))
Open Source Code No The paper is theoretical and focuses on algorithm design and proofs, with no mention of releasing source code or providing repository links.
Open Datasets No The paper describes theoretical random graph models (Erdős-Rényi, inhomogeneous random graphs) as mathematical constructs for analysis and algorithm design, not as datasets used in experiments.
Dataset Splits No The paper is theoretical and does not involve empirical evaluation with data, therefore no training/validation/test splits are discussed.
Hardware Specification No This theoretical paper focuses on algorithm design and proofs, and does not mention any specific hardware used for running experiments.
Software Dependencies No This theoretical paper does not provide details on software dependencies with specific version numbers.
Experiment Setup No This theoretical paper describes algorithms and their mathematical properties, not experimental setups with hyperparameters or training configurations.