Skip to content

Latest commit

 

History

History
58 lines (36 loc) · 2.82 KB

README.md

File metadata and controls

58 lines (36 loc) · 2.82 KB

This repository implements Momentum-Aligned contrastive Split Federated Learning (MonAcoSFL), presented at ECML PKDD 2024 - check out the poster.

It is an extension of MocoSFL, which keeps the online/momentum client models aligned during parameter synchronizations. This stabilizes the training and yields vastly improved results in deeper cut-layers that are more communication-efficient.

Comparison of MocoSFL and MonAcoSFL

image

Overview

Collaborative self-supervised learning has recently become feasible in highly distributed environments by dividing the network layers between client devices and a central server. However, state-of-the-art methods, such as MocoSFL, are optimized for network division at the initial layers, which decreases the protection of the client data and increases communication overhead. In this paper, we demonstrate that splitting depth is crucial for maintaining privacy and communication efficiency in distributed training. We also show that MocoSFL suffers from a catastrophic quality deterioration for the minimal communication overhead. As a remedy, we introduce Momentum-Aligned contrastive Split Federated Learning (MonAcoSFL), which aligns online and momentum client models during training procedure. Consequently, we achieve state-of-the-art accuracy while significantly reducing the communication overhead, making MonAcoSFL more practical in real-world scenarios.

Getting Started

Prerequisite:

Python > 3.7 with Pytorch, Torchvision

Project Structure

/run_monacosfl.py -- Source files for MonAcoSFL

/scripts -- Evaluation scripts used on a single-GPU machine

Acknowledgments

This repository is based on the original MocoSFL codebase. We would like to thank the authors for open-sourcing it.

Citation

If you find our work interesting, please cite it:

@inproceedings{przewiezlikowski2024deepcut,
author = {Przewięźlikowski, Marcin and Osial, Marcin and Zieliński, Bartosz and Śmieja, Marek},
title = {A Deep Cut Into Split Federated Self-Supervised Learning},
year = {2024},
isbn = {978-3-031-70343-0},
publisher = {Springer-Verlag},
address = {Berlin, Heidelberg},
url = {https://doi.org/10.1007/978-3-031-70344-7_26},
doi = {10.1007/978-3-031-70344-7_26},
booktitle = {Machine Learning and Knowledge Discovery in Databases. Research Track: European Conference, ECML PKDD 2024, Vilnius, Lithuania, September 9–13, 2024, Proceedings, Part II},
pages = {444–459},
numpages = {16},
keywords = {Federated learning, Self-supervised learning, Contrastive learning},
location = {Vilnius, Lithuania}
}