Exploring the Impact of Client Mobility on Decentralized Federated Learning Systems

This repository contains the code for the paper "Exploring the Impact of Client Mobility on Decentralized Federated Learning Systems".

The code is built as an extension of the DecentralizePy framework (original README below).

This work was part of the 2024 CSE3300 Research Project at TU Delft.

The code is organized as follows:

• .saved_experiments contains the results of the experiments conducted in the paper.
  • Each experiment class has its own folder, with subdirectories for each experiment run.
  • Each experiment run has its own folder for its configuration and results in [experiment*name]/machine0.
    • Configuration for the run is stored in the config.ini file
    • The initial graph used (along with its seed) is stored in a dynamic_*_.txt file
• Experiments can be run by using the tutorial directory, which the run_decentralized.sh and config.ini files
  • _submit/deploy.sh is a helper script that was used to run varying experiments with different configurations.
• src/decentralizepy contains the extension to DecentralizePy that was used for the paper. Relevant files include
  • graphs/MobilityGraph.py
  • graphs/MobilityNode.py
  • node/PeerSamplerDynamic.py
  • sharing/MobilityAwareSharing.py
• saved_figures contains figures used as well as extracted data from experiments used in the paper. These were obtained by running the plot_experiments.ipynb, compare_experiments.ipynb, and plot_graphs.ipynb notebooks.

decentralizepy

decentralizepy is a framework for running distributed applications (particularly ML) on top of arbitrary topologies (decentralized, federated, parameter server). It was primarily conceived for assessing scientific ideas on several aspects of distributed learning (communication efficiency, privacy, data heterogeneity etc.).

Setting up decentralizepy

• Fork the repository.

• Clone and enter your local repository.

• Check if you have python>=3.8.

  python --version
  

• (Optional) Create and activate a virtual environment.

  python3 -m venv [venv-name]
  source [venv-name]/bin/activate
  

• Update pip.

  pip3 install --upgrade pip
  pip install --upgrade pip
  

• On Mac M1, installing pyzmq fails with pip. Use conda.

• Install decentralizepy for development. (zsh)

  pip3 install --editable .\[dev\]
  

• Install decentralizepy for development. (bash)

  pip3 install --editable .[dev]
  

• Download CIFAR-10 using download_dataset.py.

  python download_dataset.py
  

• (Optional) Download other datasets from LEAF <https://github.com/TalwalkarLab/leaf> and place them in eval/data/.

Running the code

• Follow the tutorial in tutorial/. OR,

• Generate a new graph file with the required topology using generate_graph.py.

  python generate_graph.py --help
  

• Choose and modify one of the config files in eval/{step,epoch}_configs.

• Modify the dataset paths and addresses_filepath in the config file.

• In eval/run.sh, modify arguments as required.

• Execute eval/run.sh on all the machines simultaneously. There is a synchronization barrier mechanism at the start so that all processes start training together.

Citing

Cite us as

 @inproceedings{decentralizepy,
author = {Dhasade, Akash and Kermarrec, Anne-Marie and Pires, Rafael and Sharma, Rishi and Vujasinovic, Milos},
title = {Decentralized Learning Made Easy with DecentralizePy},
year = {2023},
isbn = {9798400700842},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3578356.3592587},
doi = {10.1145/3578356.3592587},
booktitle = {Proceedings of the 3rd Workshop on Machine Learning and Systems},
pages = {34–41},
numpages = {8},
keywords = {peer-to-peer, distributed systems, machine learning, middleware, decentralized learning, network topology},
location = {Rome, Italy},
series = {EuroMLSys '23}
}

Built with DecentralizePy

Epidemic Learning

Tutorial

tutorial/EpidemicLearning

Source files

src/node/EpidemicLearning/

Cite

Martijn de Vos, Sadegh Farhadkhani, Rachid Guerraoui, Anne-Marie Kermarrec, Rafael Pires, and Rishi Sharma. Epidemic Learning: Boosting Decentralized Learning with Randomized Communication. In Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS), 2023.

Get More for Less in Decentralized Learning Systems

Tutorial

tutorial/JWINS

Source files

src/sharing/JWINS/

Cite

Akash Dhasade, Anne-Marie Kermarrec, Rafael Pires, Rishi Sharma, Jeffrey Wigger, and Milos Vujasinovic. Get More for Less in Decentralized Learning Systems. In IEEE 43rd International Conference on Distributed Computing Systems (ICDCS), 2023.

Contributing

• isort and black are installed along with the package for code linting.

• While in the root directory of the repository, before committing the changes, please run

  black .
  isort .
  

Modules

Following are the modules of decentralizepy:

Node

• The Manager. Optimizations at process level.

Dataset

• Static

Training

• Heterogeneity. How much do I want to work?

Graph

• Static. Who are my neighbours? Topologies.

Mapping

• Naming. The globally unique ids of the processes <-> machine_id, local_rank

Sharing

• Leverage Redundancy. Privacy. Optimizations in model and data sharing.

Communication

• IPC/Network level. Compression. Privacy. Reliability

Model

• Learning Model