Jenson66 / LDPFL

Local Differential Privacy for Federated Learning

Local Differential Privacy for Federated Learning: LDPFL

This repository contains the source code of the paper titled "Local Differential Privacy for Federated Learning". For more details, you can read the paper at https://doi.org/10.1007/978-3-031-17140-6_10

If you find that this work is related to your work and it is useful, please cite our work as:

 @inproceedings{mahawaga2022local,
  title={Local Differential Privacy for Federated Learning},
  author={Mahawaga Arachchige, Pathum Chamikara and Liu, Dongxi and Camtepe, Seyit and Nepal, Surya and Grobler, Marthie and Bertok, Peter and Khalil, Ibrahim},
  booktitle={European Symposium on Research in Computer Security},
  pages={195--216},
  year={2022},
  organization={Springer}
}

The datasets

Download the SVHN dataset from http://ufldl.stanford.edu/housenumbers/

Requirements

• Keras==2.4.3
• numpy==1.23.4
• tensorflow==2.10.0
• torch==1.12.1
• matplotlib==3.3.2

Note: The python version used for the experiments: Python 3.9.4

Files

• LDPFL.ipynb: The main program that demonstrates LDPFL on the SVHN dataset. This file includes the main flow of LDPFL.
• Randomizer.py: Defines the data randomization model.
• FL.py : Defines the federated learning setup.
• Helper.py : Defines the local model and provides the functionality required for plotting the outputs related to the local model performance.

Usage

1. Install the requirements.
2. Download the dataset and save it inside a folder named "data".
3. Make sure cuda is enabled.
4. Run LDPFL.ipynb.