Skip to content
master
Switch branches/tags
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

README.md

PyTorch implementation of Multi-site fMRI Analysis Using Privacy-preserving Federated Learning and Domain Adaptation: ABIDE Results

Our manuscript is available on arxiv https://arxiv.org/pdf/2001.05647.pdf and published at Medical Image Analysis https://www.sciencedirect.com/science/article/pii/S1361841520301298.

Dependencies

  • Python 3.6
  • Pytorch 1.1.0
  • tensorboardX
  • nilearn
  • deepdish
  • numpy

Data

Data Download & Preprocessing

sh run.sh

How to run ?

Here we show a few examples using different strategies listed in the paper. Please check the meaning of configurations in each script.

Single

python single.py --split ${SPLIT} --site ${SITE}

Ensemble

python ensemble.py --split ${SPLIT} --site ${SITE}

Cross

python cross.py --trainsite ${TRAINSITE}

MIX

python mix.py --split ${SPLIT}

Vanilla Fed

vary on noise

python federated.py --split ${SPLIT} --noise ${NOISE} --type ${TYPE}

vary on pace

python federated.py --split ${SPLIT} --pace ${PACE}

Fed + MOE

python federated_MoE.py --split ${SPLIT}

Fed + Align

python federated_align.py --split ${SPLIT}

Reference

[1]Li, X., Gu, Y., Dvornek, N., Staib, L., Ventola, P. and Duncan, J.S., 2020. Multi-site fmri analysis using privacy-preserving federated learning and domain adaptation: Abide results. arXiv preprint arXiv:2001.05647.

Please cite our paper if you find this repository useful:

@article{li2020multi,
  title={Multi-site fmri analysis using privacy-preserving federated learning and domain adaptation: Abide results},
  author={Li, Xiaoxiao and Gu, Yufeng and Dvornek, Nicha and Staib, Lawrence and Ventola, Pamela and Duncan, James S},
  journal={Medical Image Analysis},
  year={2020}
}

About

impelmentation of https://arxiv.org/pdf/2001.05647.pdf

Resources

Releases

No releases published

Packages

No packages published