Reproducibility Instructions for "Differentially Private Graph Diffusion with Applications in Personalized PageRanks" Experiments

Authors： Rongzhe Wei, Eli Chien, Pan Li

This repository contains the code and instructions to reproduce the experiments presented in our paper on privacy-preserving graph diffusion framework. Please follow the steps below to set up the environment and run the experiments.

1. Create and Activate the Virtual Environment:

Open your terminal or command prompt and run the following commands to create and activate the virtual environment:

conda create -n private_graph_diffusion python=3.10.4
conda activate private_graph_diffusion

2. Install the Required Packages:

pip install -r requirements.txt

3. Running the Experiments:

The experiments can be executed by running the main.py script. The script supports various configurations through command-line arguments.

Usage:

python main.py --dataset <DATASET> --method <METHOD> [OPTIONS]

Arguments:

`--dataset`: The dataset to use for the experiment. Options are `BlogCatalog`, `Themarker`, `Flickr`.
`--method`: The method to run. Options are `our`, `pushflow`, `edgeflipping`, `all`.
`--epsilon`: The DP privacy budget of graph diffusion (default: 0.1).
`--max_iter`: The propagation iteration for PPR (default: 100).
`--beta`: The teleport probability for PPR (default: 0.8).

Example: To run the experiment using the BlogCatalog dataset with all methods, use the following command:

python main.py --dataset BlogCatalog --method all

Paper Link

📄 Differentially Private Graph Diffusion with Applications in Personalized PageRanks

BibTeX Citation

Cite our paper:

@article{wei2024differentially,
  title={Differentially private graph diffusion with applications in personalized pageranks},
  author={Wei, Rongzhe and Chien, Eli and Li, Pan},
  journal={Advances in Neural Information Processing Systems},
  year={2024}
}