totilas / DPrandomwalk

code for the paper Differentially Private Decentralized Learning with Random Walks

Differentially Private Decentralized Learning with Random Walks

Here is the code used to generate figures.

Requirements

numpy networkx matplotlib tqdm scipy

for the houses part: sklearn Pathlib typer

Datasets

The Facebook ego graph can be downloaded here: https://snap.stanford.edu/data/ego-Facebook.html and should be placed in a folder named 'facebook' in the project folder. The Housing dataset should be downloaded via the 'data' script, but can be manually retrieved there in case of problem: https://www.openml.org/d/823

Organization

• The file synthemuffcomparison.py enables to reproduce figure 1
• The folder Houses enable to reproduce the experiments of Figure 2
• The files south.py and fb.py enable to reproduce Figure 3

The code is adapted from the code of Muffliato: Peer-to-Peer Privacy Amplification for Decentralized Optimization and Averaging..